- BASIC FOR BEGINNERS JV =MI - FOUR EASY PIECES - … · omml JV =MI Jill VOLUME 1, NUMBER 2 50p...

omml

JV =MI

Jill VOLUME 1, NUMBER 2

50p U.S.$2.00

- BASIC FOR BEGINNERS- SCRAMBLE!- FOUR EASY PIECES- POWER TO THE PEOPLE

- COMPUTER IN THECLASSROOM

- PCW REVIEW: RESEARCH1978 MACHINES 380Z

Europe's first magazine for personal computers for home and business use

THE PET WITHPERSONALITY

PLUSThe International Debutof E78 - The Europa Bus

2x ObitLatches

I/O 1

Flags

Power)

'612 byre PROM

SC/MP 2 -

Data Switches1 -7--

x,-.

LEDs show Bus & Flag statusLEO dr,Irrs

1111) -

111.-111V -byte RAM

Basic 256

ar. Memory/decodingePcaAddresss 'ding

41111 Extra 512

byte RAM.

Reset &

Run modes

All component & sockets included in Basic kit except those marked ' (Full lot only)

Address & DataBusses for

Expansion

Spare

16pin skimumps Fcrrol

2+3=?Both SCRUMPI 2 and SCRUMPI 3 are powerfulMPU kits in their own rights . . . Together theymake one of the most powerful MPU Hardware/Software development systems available. Pleasewrite for details of combined kit/upgrade offer.

111.-' ALI

Soiinixa

SCRUMP) 2 is a single board MPU system based on the SC / M P2 microprocessor chip. Switches allowSingle-step/Halt/Run modes with PROM or RAM bootstraps. RAM protect and interruption. Basic kitincludes all IC sockets, all ancilliary components, SC /MP2, drivers, decoders, latches and 256 bytes ofRAM. Full kit includes additional 512 byte PROM & 512 byte RAM.

SCRUMPI 2B £55.56+VATSCRUMPI 2F £74.07+VAT

SCRUMPI 3 is a single board MPU system based on the SC / MP2 microprocessor chip and includingKeyboard, VDU interface, UART, two 8 bit parts, 128 byte RAM, 1K PROM and sockets for additional 1KPROM & 1K RAM.

SCRUMPI 3 Basic kit £1 54.92, with case & PSU £189.75.

7...VD(11 CAN DISPLAYCHARS77

EA N

KEYS

ND

pCi :44306'

ET TERS NAR:r64

E

4" INPUT pc c cow,

C4 A

Quantity discountsare available to OEMusers, Distributors,Retailers and Train-ing Establishments.

CLOCK CHIPS & KITSTYPE SPECIAL FEATURES ECHIP £KITM M5309 7 seg + BCD. RESET ZERO 4.26 8.00MM5311 7 seg + BCD 4.26 8.00M M5312 7 seg + BCD 4 DIGIT ONLY 5.65MM5313 7 seg + BCD 6 50MM5314 7 seg + BASIC CLOCK 4 26 7.00MM5315 7 seg + BCD RESET ZERO 650MM5316 Non-mpx ALARM 7.50MM5318 7 seg + BCD External digit select 4.93 8.00MM5371 ALARM. 50 Hz 12.19MM5378 CAR Clock. Crystal control. LED 9 86 14.00MM5379 CAR Clock. Crystal control. Gas discharge . . 9.86MK5025 ALARM. SNOOZE 5 60 9.00MK50395 UP/ DOWN Counter - 6 Decade 12.10 15.10MK50396 UP/ DOWN Counter - HHMMSS 12.10 15.10M K50397 UP/ DOWN Counter - MMSS.99 12.10 15.10FCM7001 ALARM SNZ. CALENDAR. 7 seg 9 00 12.50FCM7002 ALARM. SNZ. CALENDAR. BCD 9 00CT7003 ALARM. SNZ. CALENDAR. Gas discharge 9.00FCM7004 ALARM. SNZ. CALENDAR. 7 seg . . 9.00 12.50AY5. 1202 7 seg. 4 digit 4.76AY5. 1230 7 seg. ON and OFF ALARM 5.25 TBAAll above clock kits include clock PC board, clock chip, socket and CA3081driver IC. MH 15378 also includes crystal and trimmers. When ordering kit,please use prefix MHI, e.g. MHI 5309.

CLOCK MODULESLT601 Alarm Clock Module. similar to MA1002 7.00MTX1001 Transformer

707, 704, 701 0.3"DISPLAYS

727, 728, 721 0.5" (2 dig.)747, 750, 746 0.6"

1 off 1.201 off 2.601 off 1.40

MHI DISPLAY KITSMHI707/4 digit 0.3" .

MHI707/6MH1727/4 0.5"

6.00 MHI727/68.00 MH1747/4 0 6"8 00 MHI747/6

0.90

10 off 10.0010 off 11.5010 off 12.50

9 259 00

10.00

MPU SUPPORT74C00 Quad NAND 0.2474C04 Hex Inverter 0.2474C10 Triple NAND . 0.2474C42 BCD Decoder 0.9274C157 Quad Selector 2.2174C164 PISO register 1.0474C165 SIPO register 1.0474C173 3S Quad latch 0.9074C74 0.57

DM74LSOO 0.27DM74LS139 Dual 2-4 Dec 1.20DM81 LS95 3S 8 bit buff . 1.36DM81 LS96 Inv 95 1 36DM81 LS97 3S 4+4 buffer 1.36DM81LS98 Inv 97 . 1.36DM8095 3S Hex buffer 1.62DM8096 Inv 8095 .. 1.62DM8097 3S Hex Buffer 1.62DM8678 CAB Char Gen 14.30DM8678 BWF Char Gen 14.30

DM7400 Quad NAND 0.35DM7408 Quad AND 0.39DM7475 Quad LATCH 0.74DM7486 Quad EXOR 0.55CD4017 Counter 1.04CD4019 Decoder 0.54CD4040 Counter 1 04DS8833 Quad B3DI Buffer 1.99LM555 0 55MC3459 2.00

BITS & BYTESMEMORIESMM2102 1 Kx1 RAM 2.40MM21 12-2 25 6x4 RAM 3.08MM5204Q 512x8 EPROM 10.95MM2708Q 1024x8 EPROM

31.15ER3401 1024x4 EAROM 28.25MM5303 (AY -5-1013) UART 6.34

PROGRAMMED (MM5204)ETIBUG 6800 System 68 Monitor

25.95VDUBUG SC/MP System 68Monitor 25.95NIBL SC/MP BASIC (8 proms)

147.60

MPU CHIPSSC/MP 1SC/MP 11MC6800

SC/MP MPU KITSSCRUM PI 1

SCRUMPI 2BSCRUMPI 2FINTRO KITKBDKIT . .

LCDSSC1 System 68

PAYMENT TERMSCash with order, Access. Barclaycard (simply quote your number) Creditfacilities to accredited account holders. 15% handling charge on goodsordered and paid for then cancelled by customer.All prices exclude 8% VAT PLEASE SEND 30p POST AND PACKING

CATALOGUE DATA Please send SAE for catalogueXeroxed data, please phone for availability/ price

11.8910.3024.00

46.3055.5674.1066.3363.65

334.3339.84

mod

Any one or two of the above MHI display kits will interface directly with anyof the MHI clock kits

CASES (with perspex screen)VERO 1 8" x 51/2" x 3" 3.00VERO 2 6" x 31/4" x 21/4" . 3.00

SOCKETS24, 28 or 40 pin 0.60Soldercon strip skis 50 pins 0.30

BYWOOD ELECTRONICS68 Ebberns Road

Hemel HempsteadHens HP3 9QRCTel 0442 62757

[AJEJ0112130,G2JVJGJEurope's first magazine for personal computers for home and business use

CONTENTS

UK 50pUS $2.00

Vol 1, No 21978

9>

PagePUBLISHER'S LETTER 2

EDITORIAL 3

LETTERS

POWER TO THE PEOPLE W. V. RingerThis thoughtful essay reminds us that breakthroughs haveconsequences 8

A MIGHTY MICROMITE Nick ToopThe MK 14.A totally basic kit with a totally basic price. Forthose who want to get to grips with fundamentals 10

THE COMPLEAT KIT BUILDER Jeff LynchThis issue of PCW has about twentysix items. This article isone of twentysix reasons why you'll want to hoard yourcopy. 14

YES, BUT WHAT DOES THE LITTLE BEAST DO?W. G. MarshallA simple introduction to the microprocessor, using theSC/MP as a model

BASIC - THE FIRST STEPS Philip CouzensA condensed introduction for the beginner, useful as areference whenever parts of BASIC are treated in moredetail 20

TYPING WITHOUT TEARS Paul M. JessopNearly all of us type in binary - one finger or none. Here's away to go decimal and use all of our (upper) digits 24

THE MYSTERY OF THE LOST COMPUTERHarry HarrisonA great SF writer restores a bit of reality 28

EXPLOIT OR CULTIVATE? Guy KewneyThe author wields his pen against those who think thatcomputers were invented to print money 30

BOOK REVIEW 31

COMMODORE'S APPROACH TO COMPUTINGKit SpencerA personable account of CBM's marketing and designphilosophy. 32

4 THE ENTREPRENEURS PCWInterview with two pioneers of personal computing inEurope. (Photos: Y. Imamura.)

PCW CASE STUDY Boris SedaccaA small business article about a professional by a pro-fessional. (Photos: Peter McGee.)38

42

PET 2001 REVIEW John CollWhere a consultant met CBM's PET and how he took to it... 44

PCW REVIEW - RESEARCH MACHINES 380ZMike DennisA consultant gets the 380Z for a reasonable length of timeand makes a correspondingly thorough assessment. 47

FOUR EASY PIECES Neil HarrisonA virtuoso practical presentation by a consultant. But he isn'tthe man playing the piano in the illustration 54

SCRAMBLE! Graham J. TrottFirst publication anywhere of a new word game. 56

17 THE COMPUTER IS JUST WHAT THEHASN'T ORDERED Robert JohnsonLack of small computer usage is turning general practice intoa backwater; backwaters are stagnant; stagnation breedsdisease.

DOCTOR

COMPUTER IN THE CLASSROOM Charles SweetenChildren take to computing like fishes to water; or maybethat's an unfortunate analogy. Article by a consultant

THE OPEN PAGE Mike LordThe editor of the Amateur Computer Club's newsletter givethe ACC view; and other items.

EDITOR'S PLAYGROUND Patrick SuttonA timely reminder of a brilliant technique created by MaryAllen Wilkes. Article by a consultant.

61

58

64

66

E78 - THE EUROPA BUSBob Cottis, Dave Howland, Pat CroweA significant application in computing: a new bus standardmakes its international debut. This bus may turn into abandwagon. (Photos: Peter McGee.) 69

PRODUCTS PAGE 73

A VERY SMALL BUSINESS SYSTEM? Martin CollisA small software house bought a personal computer. How it DOTPIKS © K. HAMMEgot on. 36 CARTOONS 0 GIGO Er SAM

Editorial and Advertising Office:62A Westbourne Grove, London W2Phone: 01-229 5599

Publisher:A. ZgorelecEditor:Meyer N. Solomon

Policy Advisor:Peter Crofton -Sleigh, FRAS

CONTRIBUTORS:

Consultants:John Coll, Mike Dennis, Neil Harrison, Charles Sweeten,Patrick Sutton, Michael James, R. W. DavyArt:Sauveur Laurent Sant, Kathryn HammeSecretarial:Vanessa Blackburn KiddieLayout Consultants:D. Norris, T. GabosPCW Photography: Yoshi Imamura, Peter McGee

We welcome interesting articles written simply and clearly. You need not be a specialist to write for us. MS should not be more than3000 words long, lines double spaced, with wide margins. Line drawings and photographs wherever possible. Enclose a stamped self-addressed envelope if you would like your article returned.

Manufacturers, suppliers and dealers are welcome to contribute technical articles, and send product information, but we are pledgedto an independent viewpoint and will publish evaluations and reasoned criticism or praise, space permitting. Naturally there will be rightof reply. Views expressed in articles are not necessarily those of Personal Computer World.

We may make arrangements to offer our readers products at special prices, for a limited period, in line with the policy outlinedabove.

Published monthly by Irma Press, 62A Westbourne Grove, London W2. Phone: 01-229 5599. Contents fully protected by copyright. All rights reserved. Subscription rates: Britain£8 for 12 issues. Prices include postage. USA - $10 for six issues, $20 for 12 issues. Continent and elsewhere: £9.80 for six issues. Prices include postage.Printed by The Yale Press Ltd, Delga House, Carmichael Road, South Norwood, London SE25 SLY. Sole UK Distributors: Seymour Press Ltd, 334 Brixton Road, London SW9,England. Distribution to specialist shops by Intra Press.

1

PERSONAL COMPUTER WORLD

Publisher's LetterDear Reader,

It is exciting to see that we were right in thinkingthat there is enough interest in Britain and theContinent for supporting a magazine devoted topersonal computing. At many newsagents we were a"sell-out" within the first few days. The informationwe've had so far leads us to expect a sale of about25,000 copies of the first issue. This includes theremarkable number of 3,000 subscriptions to date.Quite an impressive result fora small publication.

Your support enables us to make the magazinebigger and better.

The first result is more pages and more articles.Second, we are going monthly: issue four will

appear a month after issue three.Third, we are at an advanced stage of organising

our own personal computer exhibition - inSeptember of this year. We hope that the exhibitionwill show that the personal computerist in Europehas come into his own. Many companies in thiscountry have already pledged their support and somereally important people involved in personalcomputing in America have promised to give lecturesor exhibit.

Other ideas we have will be developed and putforward in future issues.

We take this opportunity to thank some of ourAmerican friends for the hospitality and valuableadvice given to us during our visit to the UnitedStates. Nice people like Wayne Green and the staffof Kilobaud; Leslie Solomon of Popular Electronics:Stanley Veight of Computer Mart, New York; inparticular Virginia Peschke Londner, publisher ofByte: and her senior editors Chris and Blaise and thecirculation manager, Greg.

We were sorry not to have met Carl T. Helmers,the editor of Byte, who since its start in September1975 has played a major role in making Byte thebiggest and most influential magazine in the personalcomputing field - but we did manage to bring awaya photograph, published below.

FOR PCW READERS IN NORTH AMERICAFrom issue No 2 Personal Computer World will be

available at over six hundred computer andelectronics stores and news stands in the UnitedStates, Canada, Mexico and the Caribbean.

The world's leading publisher of small systemspublications, Byte Publications Inc. ofPeterborough, New Hampshire, are now our NorthAmerican distributors.

We share their belief that with them we will be ableto achieve a wider circulation in that major area andinform American readers of what is happening inBritain and the Continent.

All dealer enquiries should be addressed to:Byte Publications Inc., 70, Main Street,Peterborough, NH 03458Phone: (603) 924-7217

STAN SYSTEMA comart DISTRIBUTER

Z80 SYSTEMSOUR SPECIALITYFrom £1539 ex vat.

CROMEMCONORTH STARLEAR SIEGLERINFO 2000DYNABYTEAXIOM EX800MODEL 43 TELETYPE

31 ELPHINSTONE ROAD, HIGHCLIFFE,DORSET, BH23 5LL.Phone (04252) 77126

2


Editorial PageProgress ReportMany readers will know that PCW was first on sale atthe Microsystems '78 exhibition held in the WestCentre Hotel, London. Ours was one of the busieststands - a complete success. Then the patternrepeated itself at newsagents. Thanks to you, PCWis now on solid ground. An idea whose time hascome is rare enough; an idea that makes its own timeis even more of a rarity. PCW has not only beenwelcomed by people in computing but has alsocreated a new awareness.

With plenty of help, of course. Which leads me tohow some other magazines regard PCW! We askedsome of them to take our advertising. Reasonablyenough they nearly all refused, because PCW covers"some common ground; we'd be helping a

competitor". An exception was Electronics TodayInternational, which welcomed us in its pages. It mayor may not be a coincidence that ETI is among thelivelier publications around. As is Computing, fromwhich we reprint in this issue Dr. Bob Johnson'sarticle on the urgent need for the computer ingeneral practice. We're pleased also that GuyKewney, editor of Computing's personal computingpage, writes for us again; the man has style. Talkingof magazines, I did not mention the Americancomputing magazine Interface Age in the last issue.It now gives me genuine pleasure to recommend it toreaders interested in the American scene.Our CompetitionEntries for the PCW competition range from thegeneral to the technical and are all of them of a highstandard. The first prize is, as most of you know, aNASCOM 1. It's coming to us in kit form. Whoeverassembles and tests it will be asked to give anaccount of how he or she got on.CASBAWe're getting the first signs that the idea of theComputer Association for Small BusinessApplications is turning out to be viable as more andmore small businesses see the value of PCW as asuitable vehicle for their aspirations and objectives.The Association would be invaluable in all kinds ofways: it could put forward its views tomanufacturers; share and so lower the cost ofsoftware originated by its members; members couldgive the benefit of their expeirence to those juststarting and could even help out others if they havespare capacity; any more ideas?PCW ConsultantsOur consultants are pretty independent people -you've only to look at the articles by John Coll andMike Dennis in this issue. If I wrote an article on thePET I would praise CBM mightily for its efforts tobring computing within reach of all of us. I would besimilarly enthusiastic about any well designedpersonal computer (like Tandy's TRS80). Ourconsultants' role is to ensure that PCW is the resultof a balance of viewpoints.Take one spoonful every issueAmong the viewpoints are those of PCW readers.We take our own medicine: if we're criticised in areasoned way, and if we've the space, we print. Intelhave written us a letter - it's precise and fair.

We're thinking of having a newsletter incorporatedin PCW. It would be called Personal ConfusionWorld. It would list booboos by PCW (being a

marvellous read is not the same as being perfect) andalso amusing mistakes from other publications orpronouncements. Any contributors?New Contributors and ConsultantsHarry Harrison is a superlative writer of SF. HisStainless Steel Rat series has brought entertainmentto many thousands of SF fans, and that includes me.A new book, The Stainless Steel Rat Wants You, iscoming out this year. Philip Couzens teaches atOundle.

Dr Robert Johnson is doing a PhD at UMIST onthe real time retrieval of clinical data. He is a generalpractitioner.

Bill Ringer works for a large computermanufacturer in one of its product developmentdivisions. Pat Crowe is a senior engineer in a largeelectronics company. His homebrew system is Z80based. He's at present writing a Z80 assembler. BobCottis is a manager at the Fulmer Research Institute,doing contract research on corrosion and fatigue.He's currently building a Z80 based system using anIBM Selectric for output. Dave Howland is a seniorengineer in a large British firm and is concerned withtelecommunications. He's been using the 6800 for ayear, starting with a D1 kit, and now has a homebrewsystem.

Martin Collis is a graduate of Imperial College,London. He's a principal of Applied SystemsTechniques Ltd. which at present is mainly involvedin general commercial applications in DEC andHewlett Packard equipment. Graham J. Trott worksfor a large Government concern. He stays in late,much to the disgust of his wife, developing gamesand other programs. He's designed and built his ownM6800 system, using Eurocards.

Patrick Sutton has joined us as a consultant.He's in the Department of Physiology, UniversityCollege, London, doing a PhD in the applications ofmicroprocessors and minicomputers to medicalresearch. His article in this issue argues for therevival of elegance in computing.

Boris Sedacca is a research editor for WhichComputer? magazine, a superb publication that'smailed direct to people in computing. Nick Toop is atCambridge University. He's doing research into blackholes. Kit Spencer is CBM's man in Britain. If he hasanything to do with it the PET will become ahousehold word.

Jeff Lynch of Xitan Systems has been in dataprocessing for ten years, and in microprocessing fortwo. He's the Comart distributor on the South Coast,and specialises in marketing complete Z80 systems.

New PCW consultant Michael James trainedinitially as a mathematical physicist, and is justcompleting a PhD in pattern recognition, workingwith new parallel processing computers.

W.G. Marshall works for a large company. Heshows, in this issue, how fundamentally simple andhandy the SC/MP microprocessor is and implies thatit has been rather unjustly neglected.

New PCW consultant R.W. Davy graduated fromCambridge and Aston where he is currently a

research fellow in the Department of Engineering.He's written a cross -assembler, assembler and basiccompiler for the Intel 8008 used initially by the micro-processor group in Aston. The group's currentinterest is the Texas Instruments TI99 series ofprocessors. He was responsible for the manager forSDL's Coral 66 compiler for the 990/10.

3


LettersQuestions of Price

I first saw the PET computer at the 78 MicroSystems exhibition but I had read of it earlier in theSeptember/October 1977 issue of PersonalComputing an American magazine who made afeature of the new hardware as you are to do innumber 2 of Personal Computer World.

However there is one significant difference which isthe price quoted for this new device. In USA theprice is $595 which at the current rate of exchange of$1.95 is about £305 whereas the price asked in thiscountry is £695 which is more than twice the price inthe USA where the disposal income is very muchhigher.

Of course there are many reasons why the priceshould be different in the two countries ranging fromtechnical considerations like a different power packfor the mains in this country, different levels of taxesand transport costs. Never -the -less I would suggestthat by no stretch of the imagination could thisaccount for the discrepancy in prices. Perhaps KitSpencer or someone else from Commodore couldexplain this apparent anti-British bias.

In the same issue of Personal Computing RadioShack's announcement of TRS-80 is reported at$399.95 (£205) for the computer/keyboard or$599.95 (£308) for the full system including videomonitor and cassette recorder. Will Tandy be givinga fair deal to British computer users and potentialusers?Terence F. Hobson44 Milton Avenue, Eaton Ford,Huntingdon, Cambs, PE19 3LE

ps Why is the US price of PERSONAL COMPUTERWORLD not $1.00??PCW Because of postage rates, handling and distribution costs.

Is there a Computing Club in Aberdeenshire?Do you know if there are any clubs, organisations

or similar minded people in my area(ABERDEENSHIRE) with whom I can get in touchfor advice with my computing problems? I thinknot myself, but would appreciate any informationyou can find.R. Duncan2 Whitehills Cottages, Monymusk,I nverurie AB3 7SS

A Brickbat surmounted by a BouquetHaving only recently become interested in

personal computers, I purchased the first issue ofyour magazine (inspired by its recommendation inETI). I found, not surprisingly, that it was filled withways for me to spend my money, but not many waysin which to save it.

I am not rich, but I do have a television set, acassette recorder, and the ability to use my hands forconstruction. I also feel sure that I could find an oldelectric typewriter keyboard in Olivetti's rubbish binor somewhere. So you see, all I need now is the mostimportant item, the computer!

What I am saying is that whilst I don't expect youto recommend the construction of computers fromrubber bands, paper clips and chewing gum to hold ittogether, I do feel that you should try to bring thecomputer to the most number of people, and this canobviously be done by reducing the cost as much aspossible.

It was therefore disturbing to find that at least 5 ofyour 17 contributors had vested interests in thepersonal computer business. I don't dispute theirknowledge of the subject or the facts they present,but I take their recommendations and opinions with alarge pinch of salt. I think that if you must publisharticles by these people, it should be made clearunder the banner of the article what their interest is.

Finally, let me congratulate you on what looks likebeing an excellent publication (surprised you, didn'tit?) which I shall continue to buy.G. D. Cohen,75 Munster Rd.,London, S.W.6.PCW We do clearly state the "vested interests" of writers, in themagazine, when we give details of our contributors. And.Research Machines of Cowley Rd, Oxford, rang us just before wewent to press to announce price reductions of up to 20% on theirsystems. They must have had Mr. Cohen in mind!

Dear Sir,I have purchased and read the first issue of your

magazine PCW. I hope that you succeed in you -aims. May I as a newcomer to the computer sceneoffer some comments, (if you say no, you still get'em anyway)?

As a reader of ETI, Practical Electronics, PracticalWireless, Elektor, Wireless World, and RadioConstructor (among others), I find that the "buzz-words" of the computer fraternity confusing. As a'frinstance, I read an article and interspersed are newwords and their abbreviations i.e. UART, QWERTY,NADS, etc, etc, the explanation is in the text. So farso good. Further on in the article I come across thesame abbreviation, I've forgotten what it means so Ihunt for the explanation, by the time I've found it I'velost my place in the article, frustrating when a wholenew vocabulary has to be learned. Why not put aglossary at the end of the article in which new wordsappear? Then if a series of articles is produced thereader knows that it's the glossary to look at and nothave to search in the article.

When a new magazine that interests me appearson the market I always buy the first six issues, I thinkthat it takes that time for a new magazine to get intoits stride. I shall pay up to 1} hours pay for a goodhome computing magazine (that's roughly £1.30),but I always reckon in hours worked to buy anything

One of my other interests is Aviation, and I getmost of the magazines. If I write to Flight I always geta little card that says "Your communication of XXXXhas been received, thank you", nothing everpublished but they do known that I am alive and well.Electronic magazines readers are people who pay thecosts and profits. The only reply I had from anElectronics mag. took four months, I had forgottenwhat I had said. Another journal that I have writtento once, Practical Photography (another hobby)answered in five weeks which I think is a reasonabletime; after all, there is a magazine to publish.

So now I don't write to Electronic magazines at all,I just buy them, but there are two which I amseriously thinking of cancelling.

J. Turner,63 Millais Road,London, Eli 4HBPCW With a name like Turner and address at Millais Road, thiscould only be a graphical letter! PCW

4


Dear Sir,I enclose a booklet (A History of Manchester

Computers, S. H. Lavington, NCC Publications) I

feel you ought to have. You, and Guy Kewneywhose article I enjoyed, may like to check thisbooklet against the article and the Summary ofcomputer history on page 15, of PCW.

As a Mancunian I feel very strongly that theevidence is good supporting the claim that theWorld's first stored -program computer ran a

program successfully in Manchester in June 1948.Edsac was not running until a year later in Cambridgeand in its turn was a year ahead of correspondingAmerican developments.

So, although we may not be very good at backingand developing commercially new ideas, the newideas are there and are usually brought to a highstage of effectiveness by very few people, instead oflarge teams. In particular it is worth reading the lastsentence of paragraph one on page 32 of thebooklet.G. D. Dawson, M.Sc., M.B., Ch.B.Department of Physiology,University College of London,Gower Street,London WC1E 6BT

PCW The booklet is crammed full of interest. PCW

Intel and PCWDear Sir,

I read Guy Kewney's article "Past Procession" inyour first issue with interest and was very surprisedto note a number of serious errors. I am writing,therefore, to draw your attention to these and toprovide you with the correct information.

The story of the introduction of the first micro-computer, as described in the article, is totallyinaccurate. In 1969, a calculator manufacturerapproached Intel with a request to design a customchip set for a range of electronic calculators, and theidea of producing a computer -like architecture onsilicon chips was conceived. The result was the 4004,a four -bit word length microcomputer which dideverything that was required of it. In fact, the 4004 isstill in large-scale production for a diverse range ofapplications and, to date, very large quantities havebeen produced.

Intel went on to develop the first 8 -bit word lengthmicroprocessor which was designated 8008. Thiswas also a very successful device (contrary to yourwriter's unsupported claim), is still in production and,again, very large quantities have been produced.

Later, more powerful versions were introduced -the 4040 and the 8080. The latter device quicklybecame the industry standard microcomputer. It isdifficulty to know what the author of the articlemeans by "wretched instruction set", since he doesnot qualify or justify this statement in any waywhatsoever. Perhaps Mr Kewney might like topresent some reasoned engineering arguments toexplain what he means. His views are obviously notshared by the thousands of professional engineerswho have selected the 8080 for their design projects.The most successful hobbyist computer of all - theAltair 8800 - uses the 8080.

I am sure Motorola would be horrified to learn thatthey had developed their 6800 "by upgrading the8080". The 6800 has an entirely different architectureto the 8080 and in no way is one processor adevelopment of the other.

As a matter of interest, and to bring the story ofmicrocomputer evolution up to date, a morepowerful 8 -bit processor has been added to the Intelrange - this is the 8085. In addition, Intel have

introduced a whole family of single -chip micro-processors with up to 2K bytes of program memory,128 bytes of RAM and 27 I/O lines on one chip. Theyare the 8021, 8035, 8039, 804.8, 8049 and the 8748.Very soon we will be announcing a 16 -bit micro-computer which is similar in power to today's mini-computers.

Returning to the article; obviously, one of thereasons why humans employ computers forarithmetic is simply speed - and not precision asmentioned in the article - since humans andcomputers (even the 4 -bit 4004) can work to anyrequired degree of precision.

Any book on data base management will explainhow, in a properly organized data base, to avoid thenecessity of searching all records in file to find one atvariance with the rest. Readers interested in thissubject can do no better than obtain a copy of JamesMartin's book "Principles of Data BaseManagement" (Prentice Hall).

I would like to take this opportunity to wishPersonal Computer World every success, since thereis a need for a UK based publication of this type. It is,however, vital that the editorial staff should checkthe accuracy and validity of the material that ispublished. By all means criticize - this is an editorialprerogative - but make sure that such criticism isfair and reasoned in addition to being supported bysound engineering arguments. Rumours and hearsayhave no place in any technical journal.Keith Chapple,ManagerIntel Corporation (UK) Ltd

Microprocessors and the disabledDear Sir,

I have enjoyed reading the first issue of yourexcellent magazine. I was interested to read a shortitem on page 5 referring to the possibility of micro-processors being used in aids for the disabled, and Ithought you might be interested in a few of mycomments on this subject.

I am a teacher at a school for the physicallyhandicapped and one of the problems I often comeacross is that of the intelligent child trapped within aspastic body, unable to communicate other than bygrunts, unable to move about independently, unableto control his or her environment, and unable to learnabout the world by experimentation. Of course, aidsare available for such children, but they all suffer inthat they have evolved from equipment designed inthe '60s to answer very particular needs.

An example is the POSM (Patient operatedselector mechanism) which was a system whichallowed patients paralysed below the neck, due topolio or accidental injury, to exercise some controlover their environment. It incorporated a suck/blowtype of switch and electromechanical devices toselect and enable various mains -operatedperipherals, e.g. light, radio, TV, alarm bell, etc.Later the selector mechanism was used in con-junction with an IBM typewriter with a solenoid bankstrapped under the keyboard and with a variety ofother switches suitable for the very fine or very grossmovements produced by those afflicted by a largerange of handicaps.

There are many marvellous stories of peoplewhose lives have been transformed by the POSMtypewriter and environmental control systems, andby the telephone equipment developed in con-junction with the Post Office. There are severalsuccessful computer programmers at work todaywho had been "written off" as subnormal whenyoung children, because they could not com-municate by themselves.

5


However, when we consider whether the design ofsuch systems has kept pace with the fantastic rate ofdevelopment of microelectronics we have toconclude that the 1978 models do little more than the1968 models did, they are still designed around thesolid-state equivalents of relays and uniselectors, andthey actually cost more!

A remote controlled typewriter system now costs£1,800. The display board alone, consisting of fifty orso light bulbs fixed behind translucent perspexpanels, can cost more than a NASCOM 1! Howmuch more useful a versatile microprocessor -basedsystem would be to the handicapped child. It isexciting to visualise a disabled child with a micro-computer like Apple II, with facilities for colourgraphics, dumping onto and retrieval from cassettetape and a hard -copy print-out, all controlled by oneor two "knock switches". The system would be thechild's typewriter, note pad, reading book, picturebook, dictionary, calculator, sketch pad, gamescompendium, music synthesiser and exam paper, allrolled into one. In the not too distant future it couldliterally become the speechless child's voice.

There are but a few groups in this country workingon such systems, to help specific children or to equipparticular schools, and they can only help a minutepercentage of those who could benefit. Is itreasonable, however, to expect teachers, governorsand administrators with no knowledge or experienceof microcomputers to appreciate the opportunitiesthat are being missed? Until the educational equip-ment firms jump on the microprocessor bandwagonit is up to the individual enthusiast to try to convincethose who control the purse strings that a small boxplugged into a TV set, costing perhaps £500, is farbetter value than a well -tried and well-known device,costing three or four times as much, but comingfrom a firm with a household name.

With hardware becoming cheaper and morereadily available by the month, it is the need forspecialised software routines which is becoming themajor problem. Whilst many amateur electronicsenthusiasts will feel confident enough to construct amicrocomputer kit, thereby saving a considerableamount of money for a school, or the family of ahandicapped child, few will be able to tackle thedaunting task of writing a graphics program with thespecial editing features which may be required.

I am sure that it is here that the personal computerenthusiast will find an opportunity to make a uniquecontribution to help one of the sectors of societywhich might benefit the most from this exciting newtechnology. I feel that the bodies concerned with thephysically handicapped will be slow to acknowledgethe possible contributions which personal computingcan make, and it is from the direction of the hobbyistthat the most productive moves can be made. Head -teachers in schools for the handicapped cannot failto be impressed when they see a system runningdemonstration programs and the contagiousenthusiasm of the hobbyist is bound to be shared byall involved.

I hope that you find some of these ideasstimulating. I should be very pleased to receive anyfurther comments on this subject and would behappy to correspond with anyone who has similarinterests or problems to solve.

The eyes of the staff of our special schools mustbe opened to the opportunties afforded by this newtechnology and it is through the efforts of people likeyour readers that this can best be made to happen.Paul J. Grant,Valence School,Westerham,Kent, TN16 1QN

May I congratulate Mr. Zgorelec on his initiative inpublishing Personal Computer World, and you, sir,on the excellence of the first issue.P.B.E. Thompson,17 The Grove,Woking, Surrey.

Many congratulations on being the first in the fieldof personal computing this side of the Atlantic. Of thearticles in the first issue, it goes without saying thatDirect Addressing was the most valuable for itsinformation, and the Benchmark programs werefascinating for comparison purposes - the speedsare about two orders of magnitude better than thebest programmable calculators. (Although it wouldseem from BM48-5 that the PDP8E has negative sub-routine access time!)

Basic Pontoon was a particularly stimulatingarticle, and I now have a version of this on the TexasInstruments' TI59/PC100A. The subroutine forshuffling the cards caused some concern however,for two reasons: 1) it was time-consuming having to

look at all the cards already ordered when attemptingto choose the next card, which it would then oftenhave to reject. 2) The shuffling was not properlyrandom. For instance, if the first three cards to bechosen were aces, the last ace was still as likely to bethe fourth card as all four kings taken together. Theattached figure is my attempt to rectify these points.

The cards in the pack correspond to the numbers1-52, so that the routine may also be used in gameswhere the suit cannot be neglected.John Murrell,90 Pirniefield Place,Edinburgh.

011y's OriginalityDear Ed - I have been fascinated by the idea of

building a computer and have no experience apartfrom a two year dabble in Fortran and Cobol as ayouth. However, our immediate need is a hifi amp.Knowing nothing about electronics allows me to askwhether it would be possible to combine the two forthe sake of rationalizing the expenditure? A morerealistic question would probably be: are there anypeople interested in computers in the wilds ofNorfolk?011y Cooper,School House,18 London Rd,East Denham.

6


comartCromemco Z2 Processor

* Power of Z80 Processor Card withPower -on -jump

* S100 Bus 21 slot Backplane toconfigure the largest system

30 Amp Power Supply

Z2 = Z80 + S100

Range of Analogue & Digital Interfaces

* Powerful Disc based Software

North Star Mini -diskette System

With operating system, File & Stringhandling

BASIC

For all 8080 & Z80 S100 Bus Systems.

For greater calculation speed use theHardware Floating Point Board.

SPECIALISTS INMICROCOMPUTERS

Dynabyte Memory

16K Dynamic for Altair, Imsai, Poly 88.

16K Static with memory backing facilities,and individual 4K boundary location andwrite protect.

32K Static for large system economy.

Assembled, tested and 1 year's warrantyat list prices.

4Pift*te .7" ;;.'

tt,

Comart systems are available by direct mail order or through local suppliers:-

Xitan Systems Ltd.,31, Elphinstone Road,Highcliffe,Dorset, BH23 5LL

Phone: Highcliffe 77126

Computer Bits Ltd.,41, Vincent Street,Yeovil,Somerset.

Phone: Yeovil (0936) 26522

Our range of products is constantly being expanded - write or phone for catalogue of latestcontrollers, diskette and memory systems and options.

COMART LIMITED, P.O. BOX 2, ST. NEOTS, CAMBS. PE19 4NY TEL. 0480 74356

7


POWER TO THEPEOPLE

A view of personal computers and their possible effects on us

W. V. Ringer

How does a man boost his ego except through thethings he controls and owns? Some would questionthe value of an inflated ego and suggest that theopposite might be more desirable in the end.However, we live in a world predominantly favouringaggression, pride, nationalism, acquisition, powerand so on. The advent of the microprocessor andone of its many siblings, the personal computer, isgoing to have an unprecedented impact on all ourlives and there is every indication that the personalcomputer is going to prove to be the most farreaching extension of the person yet devised.Psychologists will say that people demonstrate theirearning and spending power through their houses,cars, holiday movies and so on in order to indicate tothe world what their personal capabilities are. Theirbookshelves and record collections are expressionsof their tastes. Will the personal computer becomethe new status symbol and will the range of collectedcomputer software and hardware be a new way ofexpressing personal capabilities and power?Whatever the consequences of the personalcomputer turn out to be, we can expect that theywon't be trivial. This article will try to show that theseconsequences could lead to: changes in publicattitudes to computers; the evolution beyond literacyand numeracy to include 'Iogicacy'; the possibility ofsafeguarding individual freedom from 'databasetyranny'; the influencing of design from networks tokeyboards; the unleashing of the creative potential ofthe masses; and finally the development of newmethods of programming to satisfy the human needto understand complex programs.

Ownership is a very powerful motivator. Thecomputer will become a personal thing, something todo with 'us' rather than something to do with 'them'.We will see computers as tools, available for all touse. The subject of computers will be demystifiedand the priests of computing, the operators,programmers and systems analysts, will simply bedoing a job professionally that we all will be doingprivately. What is it that will bring about thesefundamental changes in attitude? In all significantturning points in history there is a sudden gelling ofcertain factors that have been gathering for sometime. In the case of the personal computer, one ofthese factors is especially interesting, namely thesubtle change in emphasis of computer ownership.Owners behave differently from users. Until recently,only organisations could own computers andindividuals were merely users of the facilities. Nowthat individuals have the opportunity to owncomputers, a change can be predicted in the us -and -

8

them relationship which has hitherto existedbetween the man in the street and the computerexperts. Owning and using a computer will becomeno more unusual than owning and using a pocketcalculator, a hi-fi set or a motor car.

A characteristic of Western civilisation is the almostuniversal literacy and numeracy of its peoples. Theachievement of these skills has been helped bytechnological advances. As a consequence of themicroprocessor technology, we are now about towitness another evolutionary step forward. Ourfuture generations will not only have the ability toread, write and perform arithmetic but also toexpress relationships and solve problems throughlanguages of logic. Most commonly, everyone will befamiliar with at least one programming language.Beyond that, in the borderland of science -fiction, isthe personal computer that will interactively discussa problem and from the responses generate aprogram. Until that day arrives, however, we hadbetter learn to be programmers ourselves. Onlythrough the widespread popular understanding ofthe languages of logic can we protect our personalvalues and rights from the encroachments on thosefreedoms resulting from programmed imple-mentations of laws and procedures by narrow-minded technologists. An individual's freedom isdetermined, to some extent, by his ability to accessinformation. Libraries contain predominantlygeneralised information. There is no difficulty inaccessing this information if we wish to. Governmentdepartments, employers, banks, building societies,department stores etc., all hold personal informationabout us. The personal computer provides thephysical capability of accessing these variousdatabases. Whether or not we can persuade orrequire the database owners to allow access forexamination purposes is a matter which has yet to beresolved. The whole area of rights concerninginterrogation and exchange of information is beingtentatively approached by the government, theBritish Computer Society and other bodies under theheading of Computer Privacy. There is still a longway to go and the complexities of the problem aregrowing daily. The widespread use of personalcomputers will, no doubt, add a further degree ofurgency to its resolution.

The personal computer without access tonetworks is like a horse without a carriage. Micro-processors will facilitate the construction of'intelligent' communication networks at a variety oflevels, from local to international. Microprocessors


will also facilitate the interfacing of numerousdevices to the networks by their abilities to handlethe appropriate protocols required on the networks,independently of host computers. The interaction ofprivate, public, high-speed, low -speed, digital, voice,packet -switching and other network types is alreadycausing an urgent search for national and inter-national agreements on interfaces, speeds, protocolsetc. The entry of personal computers into thenetworks will no doubt push the need for standardseven more.Significant advancesThe market requirements of millions of personalcomputer users are bound to have an effect on thedesign and development of the most heavily usedinput-output devices. The most practical man -machine interfaces are of the screen -to -eye andfinger -to -keyboard kind. Perhaps we'll all end upwith square eyes and hammerhead fingertips.Already cathode-ray tube radiations are causinghealth concerns. This, coupled with the relativelyhigh unit cost of CRTs will force the need to developcheap, flat -screen, lower power, non -radiatingdisplays. A satisfactory alternative to the eye andfinger problem would be the use of the ear and voicefaculties instead. Significant advances have beenmade with voice recognition devices and speechsynthesizers. It is only a matter of time before theyare freely available. For the moment the keyboard ishere to stay, and this brings with it the problem ofkeyboard layout. We all wish to use keyboardsefficiently. There is nothing more irritating and timeconsumer than keying -in our creations with onefinger at the rate of one character per second afterscanning the keyboard for each character and thenscanning the source document to find the lost place.Many of us, with a bit of practice, are better thanthat, but we still envy those top-flight datapreparation operators who can key at sustained ratesin excess of 18,000 key depressions per hour. Onesolution, of course, is to learn touch-typing. Theorganisers of such courses may need to expand theirfacilities considerably. The anti-QWERTY keyboardpressure group might begin to feel desperate at thepossibility of QWERTY being set in concrete for everonce the public at large has got used to it. Or is thereperhaps another way out? New developments inkeyboard design might provide a solution to this oldproblem.

The variety of uses to which the personalcomputer can be put is limited to some extent by theavailability of ancillary equipment. The personalcomputer will follow the path already pioneered bythe mini -computers. Computer peripherals will bedesigned and manufactured to cater specifically forthe personal computer market. The limiting factor inreducing costs is that production and assembly ofelectro-mechanical components is more expensivethan the automatic mass production of large-scaleintegrated circuit chips. The technology limitations atthe moment are in the electro-mechanical field ratherthan in the logic field. The multitude of possibleapplications for the microprocessor are being heldback by the lack of equivalent advance in thedevelopment and free availability of suitable sensors,transducers etc., at the input end and electro-mechanical robotics at the output end. In the area ofthe home alone, the wish of personal computer usersto control their heating, lighting, telephone,domestic appliances and so on will require a

response from the manufacturers of electro-mechanical controls and robotic -like servo-mechanisms. A valuable spin-off will result fromhobbyists designing and building their own control

and servo systems, thereby contributing to thewealth of unfunded private research. The best ofthese developments can ultimately be exploited andproduced in volume for the benefit of all.

Is it possible that one day the pastime of televisionwatching will have to compete with the attractions ofthe personal computer? Some are concerned aboutthe effects of increasing leisure time. While many ofus are content to the passive spectators, there aremany others who would like to do somethingcreative but are not prepared to go too far from thearmchair. The program for a microprocessor whichwill control the movements of an invalid wheelchair,for instance, might be developed initially on a

personal computer. There will be scope forthousands of individuals to be creative in an abstractway for practical ends.Human ComprehensionIn the future, programs should be written like goodessays, so that they can be read and understood inan effortless way. Current program languages do notallow this because they have evolved from earlierlanguages which were designed for the convenienceof the computer. A method of writing programs isneeded which will enable the reader of a program tounderstand its purpose, objectives, limitations,exceptions and so on at all levels within the program.A program should be written with humancomprehension as its prime requirement. A programgenerator, rather than a compiler, would produce theappropriate executable machine code. The ability tocomprehend, assess and criticise a program is goingto become a very necessary skill. The proliferation ofavailable application programs for personalcomputers will bring the need for this skill forcibly toour attention. If these new programming methodsare adopted, then we will be able to interrogate aprogram structure and understand why it works inthe way it does. If we cannot do this, there is again adanger to our personal freedom. There will come aday when almost every aspect of our lives is affectedby a computer program of some kind. We will allowthis to happen because it will be very efficient. Forinstance, it would be pointless to work out manuallyeach time how many miles are left in the petrol tankwhen the on -board microprocessor presents theappropriate result continuously. We will beconditioned to trust our computers and therein liesthe danger. If we observe a program displayingundesired characteristics, we shall wish to examinethe situation in greater depth. We will require anefficient method of interrogating the very principlesand assumptions behind the logic of any program atevery level. We shall wish to do this for programsunder our control as well as having the right toquestion the logic behind the programs used bygovernment and other authorities. We are able toexamine the laws of the land and question theirimplementation. Should we not have the samefacilities regarding the programs of the land? By 1984we will surely need them!

9


A

MIC

M

THE

NICK TOOP

IntroductionI imagine that there are many people like myself whohave an interest in home computing but are put offby the cost. The MK14 kit, produced by Science ofCambridge, makes up into the cheapest completehome computer I have yet seen. The cost is less thanthat of the ICs when bought individually. The kitenables anyone proficient with a soldering iron tobuild a complete home computer. Since a voltagestabiliser is included in the kit all that one then needsis a six volt power supply capable of delivering anamp. The basic kit enables one to enter programswritten in machine code through a hexadecimalkeyboard and to show the results on a calculator typedisplay. There are also several channels for controlapplications.

Hardware and SoftwareThe kit is based on the SC/ MP2 microprocessorThis is an eight bit device made by National. The chipand all the necessary support devices are mountedon a single piece of printed circuit board which has anedge connector at each end. The entire unit is small,light, and robust enough to fit inside otherequipment which makes it ideal for experiments withsmall mobile robots and other types of controldevices.

The integral hex keyboard is of the type used in thecheaper type of calculators. It is my experience thatthis could eventually give some trouble but I do notthink the manufacturers have any alternative here. Areally reliable keyboard is a complex bit ofengineering and is correspondingly expensive. Theextra expense is probably uncalled for with thebeginner user. The keyboard is operated in a matrixfashion. There are ten columns corresponding to tenconsecutive addresses and four rows correspondingto the four upper bits of the data bus. The keysconnect between rows and columns. When a key isdepressed the bit corresponding to its row on thedata bus will go high if the address corresponding toits column is on the address lines. Since there arecontacts to both the columns and the rows on oneof the edge connectors it is a trivial matter to add amore reliable keyboard with up to 40 keys at a laterdate. It would also be very easy to experiment withother types of keyboard using touch sensitive ICs orfibre optics.

The eight digit seven segment display includesdecimal points. The eight digits correspond to eightspecial addresses and the corresponding datadetermines which segments are lit. It is thusstraightforward to display a wide range ofcharacters.

The 512 byte PROM contains a monitor programwhich enables one to examine the contents of thememory on the display and to enter new datathrough the hex keyboard. It also enables one to startexecution of a program located anywhere inmemory. While the monitor program does notsupport single stepping, which is very useful fordebugging programs, a routine for doing this isincluded in the training manual. The monitor doesnot include routines to store and reload programsusing magnetic tape although a later versionprobably will. As it stands the only way to load andrun programs is to enter them through the keyboardeach time one switches on but it is fairlystraightforward to make a simple cassette interfacewhich employs the SC/MP serial input and outputand under a pound's worth of electronics. The detailsare explained in the manual.

The basic kit has 256 bytes of RAM, and this ismore than enough for any of the programs in the

10


4

0100 ROM

0200

0304ROM

0400

0500ROM

0600

mil ROM

0800

I

0900I0

DISPLAY0A00

$

OBOO

E.TRA RAMOCOO

MOORM/I0!

DNPLAY0E00 i

RAM/I0OF00 i

BAIL RAMOFFF il

4 K

bytes

Fig 1. Addressing for the MK14

training manual. Figure 1 shows how the memory isallocated. The memory wraps round on itself after 4Kso that the RAM is directly above the PROM. Thisallows the monitor to use program relativeaddressing to save the contents of the SC/MPregisters when the user program stops execution. Ituses the locations (OFF9-OFFF). The monitorprogram has four copies in the first 2K of memory,(0000-08FF), and the basic RAM occupies the last iK,(OFOO-OFFF). The addresses for the display havesixteen copies in (0900-09FF) and sixteen more in(0D00-0DFF). If the RAM/10 chip is fitted it has fourcopies in the ranges (0800-08FF), (0A00-0AFF),(0C00-0CFF), (0E00-0EFF). The first 128 bytes ofeach copy being concerned with 10 and the last 128bytes being RAM. If the 256 byte extra RAM chip isused it is addressed in the range (0600-0BFF). Thedegeneracy of the addressing can be removed byextra gating.

The kit uses the SC /MP2 microprocessor. Thereare four addressing modes: Immediate, ProgramCounter relative (which enables one to write shortprograms which can be located anywhere in memorywithout modification). Indexed (using the addressstored in any of the three 16 bit index registers), andAutoindexed (like indexed but decrements or

increments the appropriate register before or afteruse). The auto -indexed mode is particularly useful forreading date from internal tables and forimplementing stacks to hold the arguments andreturn points for subroutine calls. There are threeuser settable flags, available through the edgeconnector, which can be used as control outputs.SC/MP has an extension register which can be usedfor the serial input and output of data or as a secondaccumulator. The edge connector also enables oneto use the two inputs. One of these is an interruptinput and can be used to make control pass to a userwritten program. It is simple to use the four 'rows' ofthe keyboard described above as four more inputs.Finally there is a delay instruction which is useful fortiming. The load from register time is about 3.5microseconds which is about average in speed. Twodisadvantages of the SC/MP chip are the somewhatrestricted instruction set and the fact that there areonly one or two byte instructions which mean thatsome operations require several steps. This results inprograms which are slightly longer than those forcomparable machines. An advantage of therestricted instruction set is that it is quicklymemorised.

ExtensionsSockets are provided for three more chips. One is fora RAM /10 device which adds 16 input/output linesand 128 bytes of RAM. The others are a further 256bytes of RAM. The RAM/10 device is a versatile chipwhich enables a program to define the sixteen 10lines as inputs or outputs. These sixteen lines can beaddressed as two eight bit words or set, reset, andtested as individual bits.

With some extra gating (there are a few unusedgates on the board) it would be possible to addressthe full 4K of memory. Only 1K of addressability isrequired for a TV type display and this would make avery desirable peripheral in conjunction with anASCII keyboard. Addressing the full 64K capability ofthe SC/MP would require rather more extensivemodifications.

ApplicationsIt is not of course possible to run BASIC or othersuch languages on a machine with this amount ofRAM. One might just manage it if one extended thememory to 4K and added a TV display and ASCIIkeyboard. Instead one has to write in machine codeor rather write in the assembler which is itsmnemonic equivalent and then do the assemblyoneself. The hex keyboard makes typing in machinecode quite quick and a program written in machinecode should run much more efficiently than itsBASIC equivalent. My personal interest incomputing has led me from writing in high levellanguages like FORTRAN to assembler and machinecode as these represent the link between theprogram as an abstract structure and the reality of theelectronics. I would therefore argue that writingprograms in assembler has certain educationaladvantages.

I have met several people who have bought anddeveloped home computer systems but seemed at aloss for applications and end up just extending them.The MK14 training manual contains several amusingand interesting programs which are well worth tryingout

If one fits the RAM /10 device a wide range ofcontrol applications arise. There will be six inputchannels, four output channels, and sixteeninput/output channels. Toys, robots, and laboratoryapplications would find the small size and power

11


consumption an attractive possibility. An interestingdomestic application would be the control of heatingand lighting since the large number of inputs make itpossible to act on the outside lighting, the roomtemperatures, the number of people in each room,etc. One could make sophisticated fire alarms orsecurity systems which operate various emergencyprocedures including phone calls. The simple way inwhich one can extend the keyboard make the kit verysuitable for games involving several players. Onecould make sophisticated alarm clocks capable ofcontrolling many independent devices and ofconditional actions. The large number of controlchannels make the kit suitable for electronic circuittesters and telephone exchanges. One could use it inmultiplexed radio control or as a PROM programmer.The list is endless.

ConclusionAs an independent consumer I was very impressedby this kit. Its low cost and versatile but economicdesign are very pleasing. I am sure that it will havegreat success.

Supply 12V ac rms at 1 amps ac rmsDisplay NSA1198 8 digit LEDR1 4K7R2 2K2R3 100KR4R5R6R7R8

1K2K21K1K1K

R9R10R11C1

C3C4C5C6C2D1D2D3D4C19C1

C2C3C4C5C6C7C8C9C10C11C12C13

C14C15C16IC17IC18

1K1K5K127pF 160v Polystyrene0.1µF 100v0.1µF 100v0.1µF 100v22 F 10v1500 F 16v1N4002

f f

not supplied

5v regulatorISP-8A/ 600 (SCAMP -11)DM745571 512 x 4 ROMDM745571 512 x 4 ROMMM2111-A 256 x 4 RAM (8111-2)MM2111-A " "MM2111-A "MM2111-A "ISP-8A/ 650 128 x 8 RAM I/ODM74LS157 Quad 2 to 1 line selectorDM74LS157 "DM80L95 Hex TriState BufferDM74LS 193 Quad Tri-State LatchDM7445 BCD to Decimal Decoder, 0/CoutputsDM7408 Quad 2 input ANDDM7408 "DM74LSO8 "DM74LSOO Quad 2 input NANDDM74LSO4 Hex Inverter

r

THE NEWBEAR COMPUTING STORE7 Bone Lane, Newbury, Berks.

Software Et Literature SectionSoftware (with paper tape) for the 6800

Games

6800 Programming for Logic DesignZ-80 Programming for Logic DesignSome Common Basic Programs

£5.95£5.95E9.95

The Bear Game E2.00 Payroll with Cost Accounting in Basic £9.95The Well Tempered Microprocessor £2.00 General Ledger System £9.953K Basic £5.00 Accounts payable and Accounts receivable E9.958K Basic £10.00Coresident Editor and Assembler E10.00 Sybexand full range for £2.80 Microprocessors: from chips to systems.

Rodnay Zaks £8.00Books p&p 50p each unless otherwise stated Microprocessor Interfacing Techniques. C207 £8.00

Some Common Basic Programs P.10 E8.00Computer designs

77-68 a 6800 Microcomputer E7.50 ScelbiSpare diagram set for 77-68 E1.50 '8080' Software Gourmet Guide Cookbook £7.25WB-1 a TTL Microcomputer E5.50 '6900' Software Gourmet Guide Cookbook £7.25Spare Diagram set for WB-1 £1.00 What to Do After you Hit Return £7.00

PCC First book of Computer Games)Zilog PCC Reference on Home Computers E4.95

Z-80 Technical Manual £3.75 Dr. Dobbs Volume 1 £10.00Z-80 P10 Technical Manual £2.25 Instant Basic E4.95

Your Home Computer £4.95Motorola My Computer Likes Me E1.65

Understanding Microprocessors E2.75130p p&pl Games with a Pocket calculator £1.75M6800 Microprocessor Programming Manual E4.50 Games, Tricks & Puzzles for a hand calculator E2.45M6800 Microprocessor Applications Manual £9.501E1.00 p&pl Best of Creative Computing Vol. 1 £5.95

Best of Creative Computing Vol. 2 £6.95M.O.S. Technology Best of Byte 03.50

KIM 1 User Manual £5.00 101 Basic Computer Games E5.256500 Programming Manual E5.00 Hobby Computers are here. £3.996500 Hardware Manual E5.00

Our range of books is constantly expanding. Please write for latest list.Adam Osbourne

Introduction to Microcomputers £5.95Vol. 0 Beginners Book E5.95 Goods are normally shipped within 24 hours subject to availability.Vol. 1 Basic Concepts £5.95 Barclay card & Access VAT at 8% for Hardware Components.Vol. 2 Some Real Products £11.95 30p postage and packing unless otherwise stated.8080 Programming for Logic Design £5.95 Cheques to be made out to 'The Newbear Computing Store'

12


THE NEWBEAR COMPUTING STORE

Hardware Components Section

Memories2102-1 £1.502102L-1 £1.602112 £3.04

4027 (25Ons) £6.002114 (450ns) £10.004116 (250ns) t.b.a.2708 (45Ons) £10.50

Please enquire for 16K dynamic and 100 up prices.

74LS0074LS0174LS0274LS0374LSO474LS0574LS0874LS0974LS1074LS11

.18

.21

.21

.21

.26

.26

.21

.21

.21

.2674LS12 .2174LS13 .5574LS14 1.26

Motorola Microcomputing IC's 74LS15 .21MC6800P £14.00 74LS20 .21MC6820P £6.20 74LS21 .26MC6850P £6.74 74LS22 .21MC6810AP £3.61 74LS26 .31MC6830L7 £11.33 74LS27 .21MC8602P £2.88 74LS28 .40

MC14536P £3.69 74LS30 .29MC3459 £2.53 74LS32 .24

74LS33 .4074LS37 .3074LS38 .3074LS40

Zilog Microcomputing IC'sZ-80 CPU 2.5MHz £15.50

74LS4274LS47

.27

.88

2-80 P10 PS £10.00 74LS48.96

Z-80 CTC PS £10.00 74LS51.96.21

Enquire for Z -80A series 74LS54 .2174LS55 .2174LS73 .34

Microprocessors 74LS74 .388080A t.b.a. 74LS75 .55SC/MP Mk II t.b.a. 74LS76 .34

74LS78 .346502 £14.93 74LS83 1.05F-8 t.b.a. 74LS86 .382650 t.b.a. 74LS107 .38

74LS109 .3874LS112 .3874LS113 .38

Interfacing IC's 74LS114 .38MC1488P £1.40 74LS125 .56MC1489P £1.40 74LS126 .5675150P75150N75154

£1.30£1.20£2.50

4F4N

5::c

74LS13274LS13674LS138

.90

.381.05

4N33 £1.95 74LS139 1.0574LS151 .9674LS153 .9674LS154 1.98

Buffers 74LS155 1.05

81LS95 £1.43 74LS156 1.05

81LS9681LS9781LS98

£1.43£1.43£1.43

74LS15774LS15874LS160

.96

.881.22

8T26 £1.84 74LS161 1.22

8T95 £1.60 74LS162 1.22

8T97 £1.60 74LS163 1.22

74367 £1.30 74LS164 1.1874LS168 2.0374LS168 2.0374LS170 2.6174LS173 2.2974LS174 1.1874LS175 1.1074LS189 2.8274LS190 2.0274LS191 2.0274LS193 1.8674LS196 1.2174LS197 1.2174LS247 .96

Standard TTL 74LS248 .967400 16 74LS249 .967402 16 74LS253 1.057403 16 74LS257 1.097404 24 74LS258 1.097406 27 74LS266 .387410 18 74LS283 .967420 18 74LS289 1.057430 18 74LS352 .967432 34 74LS353 .517437 36 74LS365 .517440 18 74LS366 .5174547472

1828

74LS36774LS368

.51

.517474 34 74LS386 .387486 34 74LS670 3.047493 577495 6574107 3474132 5274151 70 All IC's are 100% guaranteed and74153 74 manufactured by Texas, Signetic,74155 78 National etc.74157 6874163 8874164 £1 .0074175 E1.3074193 £1.3074283 E1.10

Connectors & Sockets

Subminiature D Type9 way plug9 way socket9 way cover15 way plug15 way socket15 way cover25 way plug25 way socket25 way cover37 way plug37 way socket37 way cover

.58

.68

.74

.741.09

.83

.921.50

.852.103.111.19

Edge Connectors80 way 0.1" single sided 1.8045 way 0.1" single sided 1.3043 way 0.156" double sided 4.20(for Motorola exorcisor, D1, D2kit bus)64 way plug for Eurocard toDIN 41612 1.6564 way socket for Eurocard toDIN 41612 2.9910 way plug Molex for S-50bus .3010 way socket Molex forS-50 bus .32

D -/-L Sockets (low profile gold plated)14 pin .2416 pin .2524 pin .3640 pin .6624 pin low insertion force £3.50

D -I -L plugs16 pin .6024 pin 1.20

Scotchflex connectors40 way plug40 way socket

2.402.62

Printed Circuit Boards

For: 77-68 18"x8" 0.1' edge 78 way)CPU board4K RAM boardBootstrap loader boardPrototyping p.c.b.

(low pack density)Prototyping p.c.b.

(high pack density)For: E-77 (double euro card)

4K RAM boardPrototyping p.c.b.

(low pack density)

10.0010.0010.009.95

10.91

6.257.52

For: 3U 43 way systems (44-"x£r)2K RAM board 8.50Back plane (Vero) boardapprox. 10"x4" 1.47

For: 5-50 bus4K RAM board 12.00

For: S-100 bus4K RAM board 15.00

Miscellaneous

Subminiature switchesSingle pole change over toggle.69Single pole change over(momentary action) 1.00D -I -L 8 way SPST 1.32

LED's 0.7 High LuminanceRedGreenYellowOrange

Crystals1.0000 MHz5.000 MHz10.00 MHz

Cable

.15

.18.18.18

£3.50£2.60£2.60

Computing StationeryC30 Cassettes Memorex MrX2 .75

Coding Form PadsLightweight Hexadecimal(for 6800 etc)Heavyweight Hexadecimal(for 6800 etc)Lightweight Hexadecimal(for Z-80)Heavyweight Hexadecimal(for 2-801

.80

1.10

BO

1.10

8 hole paper tape. 8' spool 1.008 hole paper tape. Fan fold. 2.00Teletypewriter paper roll 1.008' floppy discs IBM compatible

4.00

*124% VAT

Goods are normally shipped within 24 hours subject to availability.Barclay card & Access VAT at 8% for Hardware Components.30p postage and packing unless otherwise stated.Cheques to be made out to 'The Newbear Computing Store'

Send for an up-to-date catalogue to:

The Newbear Computing Store

7 Bone Lane, Newbury.Tel. 0635-46898.Callers welcome Monday to Saturday 9.00a.m. -5.30p.m.The Newbear Computing Store is a division of NewburyLaboratories Ltd.

13


H

Co

B

1

PLEAT

Jeff Lynch

1111. ERThere is nothing quite like the feeling generated onthat first power up when xxx Monitor Version 1.6,?appears on your terminal device, resulting frompatience, attention to detail and good preparation.

It's not difficult, but beware of your overwhelmingurge to "get it working".

Kits vary in difficulty - often in proportion to thequality of documentation supplied. However,indifferent documentation is no bar to a good result- it just takes longer!

1 Tools"A poor workman blames his tools", and a goodone has the correct tools to start with. Goodquality tools are an investment and the followingbeing the minimum required: -

Long Nose PliersA MultimeterWire StrippersMagnifying GlassDiagonal CuttersConductive Worksurface (earthed!)Set of Nut Drivers1 Megohm watt resistor Et about 6' wire!

Solder Removal BraidRegular PliersStanley Knife

Transistor Soldering Heat SinkSoldering Iron Stand with Sponges etc.PC Board Stand to hold 6" x 10f Boards.15-25 Watt Soldering Iron - Small andMedium Bits 60/40 PCB Solder.Screwdrivers - Various sizes including

14

Philips Head.PCB Cleaning Solvent, (RadioSpares Cat No.555-134), Cloth, Brush.

2. General Hints and Tips(a) Read ALL the documentation provided. Read

it ALL again. Find the errata sheets and ifthere aren't any ring your supplier and check.Amend instructions as required by the erratasheets ensuring that the changed instructionsstill make sense.

(b) Work in a clean, well lit area, keeping all thecomponents together. Check the componentsagainst the modified parts list (using theerrata sheet) before assembling anything. Tryto avoid working in a distracting environment.

(c) Be patient. Work only when you cancomplete the section you start in one sittingwithout getting tired. You will need all yourconcentration to sort out those capacitorsetc., so don't start installing them atmidnight. Check every move as you make it. Ifyou have any doubts, check it again beforesoldering. Not only will you have to correct amistake (which takes time) but you will haveto find that mistake. A 16K memory board hasover 2,000 solder joints, and three days afterstarting you'll be at your wits end trying toestablish which one has got the wrongcomponent in it!

(d) Check voltages supplied to IC sockets beforeinserting ICs are correct, and within thetolerances noted by the manufacturer.

2. (e) Never install or remove ICs in sockets withpower on the board, nor insert or remove


boards from the chassis/mother board withpower on the machine. Basically ensurepower is OFF before doing anything to theinterior of the machine

(f) Always install CMOS or MOS IC/Devices lastfollowing the procedures outlined in part 4.

3. P.C. Board Assembly(a) Examine the bare board for shorts between

traces and IC socket pin positions etc., onboth the component and solder (front andback!) side of the board. On S100 boards,check for zero resistance (shorts) between -

Pin 1 (+ 8v) Et 50 (Ground)Pin 21+ 16v) Et- 50 (Ground)Pin 51 (+ 8v) a 50 (Ground)Pin 521- 16v) Er 50 (Ground)

Et Voltage Regulator Output pins Et Pin50If a short exists use a magnifying glass toexamine the board. Look in areas of highpin/trace density such as IC socketpositions, and also search especially carefullyunder silk screen markings. Take your timeexamining the board.

(b) Component Device IdentificationFollow instructions carefully. Use a colourcode chart to identify resistors etc., find afriend to help you with the capacitors. I findthese the most difficult, but normally they canbe identified by a process of elimination bysorting them into piles of differentvalues/shapes and comparing the number ineach pile to the parts list. Remember that100N is 0-1 Mfd, 10N is 0-01 Mfd. Withcapacitors, take extra care to get the polaritycorrect where necessary. (The +ve lead isidentified by: - the lead with a single smalldot nearest it, ii) the lead nearest a + markedon the component, or iii) if the componenthas a single large dot on it, the +ve lead is tothe right when the dot is facing you, and theleads are at the bottom).

(c) InstallationFollow the manufacturers instructions in theorder of assembly. Bend leads on diodes andsmall capacitors carefully, using the longnosed pliers to hold the lead near thecomponent, thus avoiding undue stress onthe component. When it comes to installingIC sockets, it is best to either, put a strip ofstiff card across a row of sockets, and tape ittightly so the sockets won't lift off the boardwhen it is turned over for soldering, or ii) oneach socket blend diagonally opposing pins tohold the socket tight against the board.

Make sure that sockets are seated firmlyagainst the board before soldering. Install asection of components at a time, bending theleads to keep them in place. Tick them off theparts list as you do so on completion of a set(e.g. all the resistors) make sure you havenone left.

Do not install switches until after you havecleaned the board (see Part 5.).

(d) Solderingi) Use a low wattage iron (15-25 watts), and

60/40 Rosin core solder (Not acid core).ii) Keep the bit clean using a damp sponge.iii) Solder only on the non -component side

unless told to do otherwise.iv) Check before soldering that it is the right

component in the right place.v) Apply the bit to both the component lead

AND the board foil, and then apply the

solder to the bit.vi) It should melt immediately, and flow

smoothly into and around the hole/lead.Less than solder is normally sufficient.

vii) If the solder does not flow - remove thebit immediately - the junction you aresoldering is probably dirty or corroded.Clean it with a little solvent (away fromthe bit! or flames etc) before retrying.

viii Remove the bit as soon as the solder hasflowed into the hole - leave to coolbefore going on to the next joint.Disturbing the joint before it cools willcause "cold" solder joints. Do not keepthe heat applied for more than a secondor so as overheating can i) cause thecopper foil to lift and ii) cause "heatshock" to the components you aresoldering. Both of these faults will causelong term reliability problems.

ix) The joint should be bright and shiny witha smooth surface. A dull or rough surfaceis an indication of a "cold" joint. Re -applyheat in these cases.

x) Once all component have been soldered(except any switches) all unfilled holesshould be solder filled. Although they are"through plated" this is no guaranteethere is any electrical or mechanicalcontact between one side and the other.Solder filling is the only safe method.(The through plating is there to provide asurface for the solder, according tomilitary specifications!)

xi) Cut off excess leads using diagonalcutters.

(e) DesolderingIf you've been extra careful, you'll probablynever need to desolder, but you might have toremove a duff component.i) Never just heat and pull. You will do

irreparable damage this way, with thecopper foil lifting off, the throughplatingpulling out etc.

ii) Apply desoldering braid to the joint, andthen apply the bit to the braid. Don't pushtoo hard - as the braid acts throughcapillary action and needs the gapsbetween individual wires to work. Whenthe braid has sucked up some solder,remove bit and braid simultaneously. Cutoff used braid, clean the bit and repeatuntil you have removed as much solder aspossible.

iii) Be careful not to overheat!iv) When 95-100% of the solder has been

removed, heat the lead andsimultaneously use a screwdriver tocentre the lead. Gently pull the lead fromthe hole while applying heat.Do all this in short stages, allowing thecomponents to cool between attempts.

4. Handling MOS/CMOS DevicesMOS/CMOS devices are sensitive to static voltagedischarges, and are readily damaged by staticdischarged from handling during assembly.However, following the rules below will make thishighly unlikely - and most manufacturers insistthat these sort of rules are followed to comply withwarranty requirements.(a) Using the 1 Megohm resistor in the "Tools"

list and the 6' wire, connect yourself, via theresistor, to a good ground (cold water pipe,radiators etc). The wire can be attached to a

15


metal watch strap, or just noosed round yourwrist. The resistor is a safety precaution tolimit the extent of any accidental shocks youmay get by touching "live" wires!

(b) Wear cotton clothes - synthetics are badnews!

(c) Spread out a conductive work surface andearth that as well. A sheet of metal is best, buta strip of Aluminium cooking foil is a goodalternative. Do not use any plastic or glass.They are not acceptable.

(d) Try to work in a medium humidityatmosphere. Dry and/or cold air assists thebuilding up of static and should be avoided ifpossible.

(e) If you have to solder a MOS/CMOS deviceensure that the soldering iron tip is grounded.

(f) If you use any tools, ground them beforetouching the device - insulated pliers forinstance may have picked up a charge whenyou last used them.

(g) Basically, try to keep everything in contactwith everything else - and all of it grounded.

5. PC Board Cleaning and InspectionWhen all the components except switches havebeen soldered - the Board should be cleaned.This step is vital. Multicore solders tend to spitrosin and minute drops of solder around the board.This has all got to be removed. A good qualitysolvent should be obtained (RadioSpares No. 555-134 is good as well as being non flammable) butindustrial alcohol will do at a pinch. You will alsoneed a stiff r brush.(a) Liberally pour the solvent onto the solder side

of the board. "Scrub" the solvent off. Pouron some more solvent and repeat. Make sure

you cover the whole board. When satisfiedyou've covered the whole board thoroughly,blot the remaining solvent up with a cloth.Finish off off by wiping the contact fingersclean with a solvent soaked cloth.

(b) Now, using a magnifying glass, inspect theentire board for solder splashes, shorts, coldjoints, unsoldered pins.

(c) When you're satisfied - carefully solder anyswitches into position, trying to avoid solderspits etc. Switches and solvent don't gotogether well - the switches tend to dissolve!That's why you should leave them to the last.

6. Inserting IC's into SocketsThis is not a difficult job, but there are two thingsto consider. Use your commonsense and check ICorientation before insertion.(a) Line the pins of the IC's up as accurately as

possible before you try to insert the IC.(b) Make sure you push hard to get the IC to seat

correctly. In some cases it will appear torequire too much effort. Don't worry. Thesockets are meant to grip tight.

(c) Once in, check that no pins have been bent upunder the IC. It should be obvious if you lookend on at the IC.

7. SUMMARYIf you haven't built a kit before, don't worryovermuch. Just consider the value trebled andthink professionally! Be patient. Read these tipsagain. Read the instructions with the kit until youunderstand them. It's not as difficult as I may havemade it appear, but if you follow these rules, you'llprobably end up with a professional result whichwill work first time and give long useful service.

electronicstigg

Home Computing SupplementSURVEY: We take a look at what is available, in personaland development systems for the enthusiast.MEMORIES: CCD, bubble, toil and trouble, PROMs,EROMs, floppies and core - memories explained.GAMES: Everybody likes to play games, why play withyourself when you can play with your computer?

PERIPHERALS: Printers, VDUs, discs, cassettes, with-out them your MPU is an exceptionally dumb machine.LANGUAGES: BASIC to APL what are the benefits ofhigh level languages, and how to use them.GLOSSARY: Even in this short advertisement we haveused terms that a lot of people do not know, A thru Zee.

OTHER FEATURES INCLUDE: LED Spectrum Analyser project, Analogue 'Mastermind'project, FM/AM Radio project, QUARKS for beginners, plus all the regular features - toonumerous to list! Available at all good newsagents, don't miss it.

All in the JUNE issue, on sale MAY 5th 45p16


YES, BUT WHAT DOES

THE LITTLE BEAST DO?W. G. Marshall

Introduction

The last couple of years or so have seen the arrival ofa large range of microprocessor 'chips' and theirintroductory kits. Most people have probably seenthe adverts for them by now and wondered just whatit is they do and how they do it. In this article I intendto describe the basic functional blocks which makeup a microprocessor system, with particularreference to the National SC/MP. In a small set-upconsisting of processor chip, read-write memory(RAM) read-only memory (ROM) and a few controllogic gates, simple program execution should beeasy to follow. Assuming that the reader knowssomething about the binary number notation andthat a decimal number is represented by a string of'Vs and '0's, it can be seen that in an electricalsystem a 'high' voltage is a '1' and a 'low'(2:..-0v) is a '0'. Theoretical binary can thus beprocessed in the machine as a series of high and lowvoltage levels.

Inside the MicroprocessorFig. (1) illustrates the essential working blocks of theprocessor chip, greatly simplified, as it is notnecessary to understand all the inner workings inorder to use these little beasts. The timing andcontrol logic determine the movement of data andthe operations that are carried out upon it. Theprocessor operates in a series of cycles, each ofwhich consists of two parts: instruction fetchfollowed by instruction execution. This cyclicoperation is common to virtually all digital computersincluding all microprocessors. The working registersin the processor can be thought of as groups ofbistable flip-flops, each flip-flop being able to store a1 or 0 logic state. In the case of the SC/MP, theProgram Counter (PC) and Pointer registers eachconsist of a row of 16 flip-flops. The Accumulator,Extension and Status registers each contain eight.The term 'Bit' is given to a single unit of data, and 8

TIMINGLOGIC

CONTROLLOGIC

ADDRESS LINES

READ/WRITE STROBE

< DATA LINES >PROGRAM COUNTER PC

POINTER REGISTER P1

POINTER REGISTER P2

POINTER REGISTER P3

ALU

EXTENSION

STATUS

ACCUMULATOR

MEMORY(RAM)

FIG. 1. SC/MP SYSTEM BLOCK DIAGRAM

bits are usually called a 'Byte'. Hence an example ofa byte is: 01010010, which has the decimal equivalentof +62. Now we can say that the accumulator is 8bits long and the program counter is 16 bits in length.

Taking the program counter first, its outputs goout on the address lines of the microprocessor(MPU) at instruction fetch time, and so its contentsdetermine which instruction is read from thememory. The instruction is read into the MPU fromthe 8 data lines. Having received its instruction, thecomputer now proceeds to obey it. It may forexample, have to read into the accumulator registerthe contents of some memory location. Whenexecution is complete, the program counter isincremented by one and a new computer cyclebegins with another instruction fetch.

Most computers have one or more registers calledaccumulators. The SC/MP has two (the extensionregister is virtually an accumulator), the 6800 has twoand the 8080 one. Most arithmetic and logicaloperations are carried out on data contained in theaccumulator, and its tends to be the register thatsees the most use.

Next come the pointer or index registers. Thesemay be loaded by program and their contents may beselected, also by program, to go out on the addresslines at instruction execute time. Hence they addressthe store location where program data will be storedor taken from.

The function of the various bits in the statusregister varies from MPU to MPU, but some bitfunctions are fairly standard. One bit is normallydefined as the Carry/Link bit, and is set high whenthe result of an arithmetic operation causes a carryfrom the most significant end of the accumulator.This is useful when handling numbers of greaterlength than 8 bits. For instance if two 16 bit numbersneed to be added, the least significant (rightmost) 8bits of each are added first. The result is stored, andthe most significant (leftmost) 8 bits of each addedtogether, the carry/link being added in as well topropagate the carry.

Arithmetic in computers is normally carried outwith binary numbers in what is called '2'scomplement' form, which results in the mostsignificant bit in a byte being the sign bit. That is, if itis set to a 1, then the number is negative, if not setthen the number is positive. It isn't quite that simplehowever! For example: -00010110 is +22 in decimal

sign bit

To get -22 invert (change Os to is and vice versa) allthe bits in the byte and add 1 to the result: -11101001

+1

11101010 is -22 in decimal

Now it can be seen that with an 8 bit accumulator,numbers in the range -128 to +127 may be handled.But what happens when you add two numberstogether when each lies close to the same end of the

17


range? For instance add +120 and +112: -01111000 = 120

+ 01110000= 11211101000 is -24 in decimal?!

The answer is -24 in 2's complement form andillustrates an overflow error. If the sign bit of theresult is different from either of those in the two databytes then an overflow error has occurred. A bit inthe status register is normally set to indicate thiscondition, and the program may by written to checkthis bit after arithmetic operations.

That about covers the internal registers and wemove on to the block in Fig. (1) marked ALU orArithmetic Logic Unit. This is simply a piece of logicthat performs the addition, complementing, etc.functions on two data bytes, usually leaving theresult in the accumulator. The function it carries outis determined by the current instruction beingobeyed.

The Computer MemoryThe memory of a computer has two tasks, one tohold the program or list of instructions to beperformed by the MPU, and the other to supply orreceive data as required for the program execution.In an 8 bit system, each location in the memoryaddressed by the MPU contains 8 bits of data.Sometimes 8 bits are not enough to hold a completeinstruction and so two or even three consecutivelocations are used to hold one machine instruction.The SC/MP uses one and two byte instructions andthe handling of these will be shown by detailedexamination of a simple program. As well as theaddress and data lines going to the memory from theMPU, there is a control signal which determineswhether data currently on the data lines from theMPU is stored in the addressed location, or whetherthe RAM is required to supply data. This is usuallycalled the Read/Write Strobe. A read-only memory(ROM) does not of course require this signal as it isonly possible to read data or program from such adevice. At this point I should mention that the groupsof data and address lines are generally called 'Buses'.Hence a 6800 MPU is said to use an 8 bit data busand 16 bit address bus. If you examine the chip youwill find 8 pins allocated for the data wires, and 16 forthe address wires.

Examination of a Simple ProgramOne of the best ways of getting to know how a

computer works is to play with one! However if thatis not possible yet, consider the following smallSC/MP program which adds one number to anotherseveral times and then halts.

MemoryAddress

MnemonicInstruction

Code or data1st Byte 2nd ByteBinary Code or data

1, 2 LD 0(P1) 11000001 000000003 CCL 000000104, 5 ADD 11P1) 11110001 000000016, 7 ST 1(P1) 11001001 000000018, 9 DLD 2(P1) 10111001 0000001010,11 JNZ -11(PC) 10011100 11110101*12 H LT 0000000013 +2 0000001014 +3 } ** 0000001115 +4 00000100

The above is a small section of program. (Lightly tickoff the rows in pencil as you read).

*Leftmost bit is 1, so this is a negative number storedin 2's complement form, = -11 in decimal.**Initial values held in locations 13, 14, 15.

The program assumes that the pointer register P1 isset to 13 (i.e. data to be manipulated is stored fromlocation 13 onward) and that the accumulatorcontains zero on entry into the program. The columnon the left of the listing gives the address in decimal,of each instruction or piece of data. The middlecolumn contains a mnemonic representation of theinstruction or data. This is easier for the programmerto follow than the binary codes, in the final column,which the MPU understands. This binary code is theform in which instructions and data are stored in thememory. You can see that most instructions are ofthe two byte variety, requiring two consecutivememory locations to hold them.

On starting the program, the program counterregister in the MPU is at 1 and so during theinstruction fetch part of the cycle the processor readsthe instruction contained in location 1 of thememory. This is achieved because 1 goes out on theMPU address bus and the memory puts the contentsof location 1 onto the data bus. This instruction islatched into the MPU and it decodes the binary codeof the instruction to find out what it has to do next.The most significant bit of the byte is set to 1, so theMPU knows that it only has half the instruction. Theprogram counter is incremented to 2 and the byte inthat location (2) read in. The MPU now proceeds toexecute the double -byte instruction which says:"Load into the accumulator the data which is

contained in the memory location addressed bypointer register P1 (ie 13), offset by 0."

Quite a mouthful! It should be easy to follow,except for the: "offset by 0". This is an additionaladdressing feature available on most MPUs wherebythe address that goes out on the address bus atexecute time is the arithmetic sum of the pointerregister contents and the second byte of theinstruction. In this instance the second byte or offsetis zero, so the address 13 (in P1) goes out on theaddress bus. The data contained in location 13comes back on the data bus and ends up in theaccumulator.

The first computer cycle is now complete, theprogram counter is incremented to 3 and the next(single byte this time) instruction is taken from store.Bit 8 of the instruction is zero this time, so the MPUknows that it has the complete code and executionfollows. This time all that is done is to set thecarry/link bit in the status register to O.

The next instruction (in 4,5) adds the contents oflocation 14 to the accumulator. The data is takenfrom location 14 this time because we do have anoffset of +1, this time against P1 which is set to 13.

The instruction in locations 6 Er 7 stores the resultof this addition, ie. the contents of the accumulator,in location 14, as the offset relative to P1 is again +1.

The next instruction in locations 8 8- 9 is a

composite one and in the execute part of the cycleperforms three operations. First it reads the contentsof location 15 (note the offset of +2) into theaccumulator, subtracts 1 from the latter and thenstores the accumulator contents back in location 15.So now we have +3 in that location.

The Decision TimeThe next operation (in 10,11) is an example of theinstruction type that separates the computer fromthe average pocket calculator. This is a decisionmaking instruction, which depends on the contentsof the accumulator. Remember that the latter nowcontains the same data that is held in location 15. Ifthe accumulator is zero then the program continueswith the instruction in location 12. If it is not zero,

18


then the PC (which contains 111 is reset back by theamount (= -11) indicated by the offset. That is, to 0.The PC then increments to 1 and the LD 01P1) codeis fetched and executed again. The program thuscontains a loop' with a 'loop counter' in location 15,the latter decremented by 1 every time the computercycles round the loop. Eventually location 15contains 0 and so does the accumulator - and thedecision now is that the computer executes theinstruction in location 12 instead of returning to 1.This is a Halt command and the computer stops anyfurther operation.

Let us now examine the contents of the three datalocations and see what they contain. Location 13should still contain +2 as no operation other thanreading (which doesn't alter the contents of memory)has been performed upon it. Location 14 containsthe 'answer' or final result of the calculation. Usingthe initial data values given, this should be +11.Work through the program by 'hand' to check thatthis is so, and that the sequence of operationsperformed is: 2+3+2+2+2. Location 15 will be 0 as itis decremented for each trip round the loop ofprogram steps.

ConclusionIn a short article such as this, the immense versatilityof the computer or MPU cannot be fully explored.However I hope the reader's appetite has beenwhetted and that he will wish to investigate furtherthe computer's power of decision making, byassembling one of the many MPU kits on the market.Even the simple ones may form the basis of moreambitious projects in the future.

Mol MINE OF INFORMATION LTD. MolChapman & Hall

INTRODUCTION TOMICROCOMPUTERS

are pleased to offer you three exceptional books byAdam Osborne Et Associates of California, USA.

Each volume can be read alone as a self-containedbook. The set of three volumes spans the gap betweenthe layman and the specialist.

VOLUME ZERO 'The Beginners Book' is clearly andsensibly written with pictures and diagrams whereverthey might help a beginner £5.90VOLUME ONE 'Basic Concepts' provides sympatheticteaching leading to a thorough understanding of con -cepts common to all microcomputers £5.90VOLUME TWO 'Some Real Products' is the handbookfor the professionals which describes, compares and con -trasts twenty leading microprocessors in detail . . £11.90

BYTEOne year's subscription to BYTE magazine £14.90

Prices include postage. To pay by ACCESS give card number, name,

address and signature.

MINE OF INFORMATION LTD.1 FRANCIS AVE, ST. ALBANS AL36BLTel. 0727-52801

The scientific, technical and medical divisionof Associated Book Publishers Ltd,11 New Fetter Lane, London EC4P 4EE

SMALL SYSTEMSCOMPUTERWe

SOU RCEBOOKEDITED BYJ C BOONHAMThis text, the first of its kind, will be an invaluable aidto all who are interested in the purchase, constructionand usage of mini computers. It is specificallydesigned to assist the prospective purchase in manyfields, i.e. home enthusiasts, educational and smallbusiness users, engineers and others with scientificor commercial applications.

The main sections of the book list comprehensivelythe available equipment and specifications and pro-vide references to suppliers and distributors ofhardware, software and ancillary equipment. Everyeffort has been made to accommodate the wideranging and complex nature of the subject by intro-ducing newcomers to computing to the essentialelements and special terminology of the subject.

Main Contents Headings: Section 1 : Availablehardware; Section 2: Software; Section 3: Costaspects of hardware and software ; Section 4 :Manufacturers and suppliers; Section 5: Glossary ofterms; Section 6: Appendices; Section 7: Biblio-graphy. Index.

10 May 160 pages IllustratedPaperback 0 412 1 5680 6 £6.50Distributed by Chapman & Hall onbehalf of THE ROCOCO PRESSThe above price is net in the UK only.

19


AS ICTHE FIRST STEPS

Philip CouzensStep after step the ladder is ascendedGeorge Herbert, Outlandish Proverbs (1640)

BASIC, Beginners' All-purpose Symbolic Instruction Code, was originally designed at DartmouthCollege, U.S.A. to be a simple introductory language from which students would progress to themore powerful high level languages such as FORTRAN, COBOL, ALGOL and PL/I. The grammaticalrules of BASIC are so few and simple that the user can quickly program the computer to solve hisproblems.

Although BASIC is so simple, it is still a very usefuland powerful high level language suited to bothmathematical and business situations alike. Due toits simplicity, interpreters and compilers for thelanguage are small, making BASIC particularlysuitable for microprocessor based systems in whichcore storage may be limited.

If one has experience of one of the popular highlevel languages, particularly FORTRAN, thenconverting to BASIC should take only a few hours.The remainder of this article is intended for thosewith no previous programming experience, but mightalso be useful for those wishing to change fromanother language.

BEFORE PROGRAMMINGBefore programming can begin, the system has to beswitched on and the BASIC interpreter has to beloaded into memory. The method of doing this variesconsiderably from system to system, obviouslyinstructions for doing this cannot be given here. Itwill be assumed that the system is ready and workingin BASIC.

At least two of the control keys should then beidentified. These are the 'carriage return' key whichis pressed every time something is entered at thekeyboard, and the 'escape' key which is used to stopthe computer doing a program. If no particular key ismarked 'escape', then the instruction is probablygiven by holding down the 'control' key whilepressing another key - check in your manual forthis.

One further key that is useful is the 'rubout' or'delete' key (sometimes back -arrow or underline).This has the effect of erasing the last character whichwas typed. Pressing it twice will erase the last twocharacters, and so on.

WRITING A PROGRAMAny line which is now typed in at the keyboard willbe treated as a command, and this will be executedwhen the 'carriage return' key is pressed. If a linebegins with a number (any integer between 1 and10000), then that line will be stored as part of aprogram.

So, for example, the first line of a program is typedin:

10 PRINT "HELLO"

This, of course, is followed by a carriage return.*Every program in BASIC should be terminated bythe instruction END, so the last line is typed in:

20 END

There which will be:yped out at the terminal if the command LIST isssued at the keyboard; the lines of the program areyped out in order of their line numbers, not in the)rder in which they were typed originally. To executehe program the command RUN is given. This)rogram simply prints H ELLO and then stops.

If it becomes necessary to insert further lines into-le program, then this is done by choosing a suitable

line number which has not already been used. Forexample, a line is inserted between the two existinglines:

15 PRINT "HELLO AGAIN"

If the command LIST is now given, the wholeprogram is displayed at the terminal:

10 PRINT "HELLO"15 PRINT "HELLO AGAIN"20 END

*This is not necessary on all systems, but it is usefulto conform to standards.

If it becomes necessary to change a line, then thisis done by creating a new line with the same numberas the one to be replaced and retyping the whole line.A line is deleted by typing in the line number followedby 'carriage return'.

So it is seen how a program is built up. It is usualto choose 10 for the first line number and to go up intens, that is 10, 20, 30 . . . This leaves plenty ofunused numbers so that lines can be inserted ifrequired. The command RENUMBER, which issupported by most systems, changes the linenumbers to fit this pattern.

When it is wished to lose the existing program and

20


start a new one, the instruction *NEW is given. Theold program is lost completely and a new one may becommenced.

The important BASIC commands which we havemet so far are:

(i) RUN - the current BASIC program isexecuted

(ii) LIST the current BASIC program isdisplayed at the terminal

(iii) NEW the current BASIC program isdeleted so that a new one maybe written

(iv) RENUMBER - the lines are renumbered to fitthe pattern 10, 20, 30 . . .

ARITHMETIC IN BASICNumbers are expressed in BASIC in three ways:

(a) Integers - written without a decimal point,e.g. 4395

(b) Decimals - written with a decimal point, e.g.31.416

(c) Numbers written in standard form. 1 08 x 10-4= 0.000108 is written as 1.08E-4

A negative number is preceded by a minus sign inthe usual way, for example -4395, -31.416,-1.08E-4.

These notations can be used as and whenconvenient.

*On some systems the command is SCRATCH

Numbers will be typed out by BASIC in the mostsuitable form, the user is normally not able tochoose. Unless a number is particularly large or smallthe integer or decimal form will be chosen.

All arithmetic expressions have to be typed in atthe keyboard, and therefore the notation is limited bythe facilities of the keyboard. Addition andsubtraction are denoted by the usual signs + and -.Multiplication is denoted by the sign * and division bythe symbol /. The up-arrow 4 is used forexponentiation (raising a number to a power), andbrackets are used as in ordinary arithmetic.

As an example, the expression

(3.14 + 2 xi)492 - 4.6 x 106

could be evaluated

in BASIC by running the program:

10 PRINT (3.14 + 2°1 /2) 44/ (9 4 2- 4.6E6)20 END

BASIC uses standard algebraic hierarchy whenperforming arithmetic, the order of operations beinggiven by BEDMAS (brackets, exponentiation,division = multiplication, addition = subtraction).Operations of equal priority are performed from leftto right. So the answer to 1 + 2* 3 is given as 7 sincethe multiplication is done before the addition. Noticethat 12/3*4 is given as 16. The order of performingoperations is in fact exactly the same as one woulduse in ordinary arithmetic.

LIBRARY FUNCTIONSAll BASIC systems (except those called Tiny Basicetc) should support at least the following set of

library functions:

FunctionSIN (x)

COS (x)

TAN (x)

ATN (x)

SQR (x)LOG (x)EXP (x)ABS (x)

INT (x)

SGN (x)

RND (x)

Meaningthe sine of x, where x is an angle inradiansthe cosine of x, where x is an angle inradiansthe tangent of x, where x is an angle inradiansthe principal angle (measured inradians) whose tangent is xthe square root of xthe natural logarithm of xthe value of ethe absolute value of x, ABS (-1.9) =1.9, ABS (1.9) = 1.9the largest integer not greater than x,INT (6.3) = 6 INT (-6.3) = -7the sign of x, has value 1 if x is positive,0 if x is 0 and -1 if x is negativex is positive, 0 if x is 0 and-1 if x is negativea random number between 0 and x

In the above, x, called the argument of the function,represents an expression of any complexity, and mayinclude other functions. Note that all angles aremeasured in radians so that degrees have to beconverted (multiply by rt/ 180 = 0.01745331.

It should now be possible to use BASIC as ascientific calculator.

EXAMPLES.

(1) Evaluate sin 30° + \/210 PRINT SIN (30 * 0.01745) + SQR (2)20 END

(2) Evaluate (1n 3 + e v )3 4

10 PRINT 2/3 * (LOG (3) + EXP (SQR(2))) 320 END

Note that several levels of brackets are allowed, butone should be careful that they are closed correctly.

VARIABLES IN BASICVariables are used in BASIC in a way not dissimilar tothat of ordinary algebra. The equals sign, however, isused in a very different way and this is stressed bythe use of the word *LET. For example, consider theBASIC statement

10 LET A = 4.2

This assigns a value of 4.2 to the variable A. Themeaning is made clear by reading the instruction as'let A become 4.2'. The expression on the right handside of the 'becomes' sign can be of any complexity,and can contain any number of variables (includingthe variable on the left hand side). The left hand sidemust contain only one variable.

Consider the statement 30 LET A = A + 1.

Although this would be nonsense in mathematics, itis allowed in BASIC, meaning 'let A become what itwas plus one' and has the effect of increasing A byone. It is, indeed, a very useful and much usedfacility.

*In many versions of BASIC the word LET may beleft out.

Ordinary variables which represent numbers inBASIC can be given any of the 286 names which

21


consist of a single letter: A, B, C, . . . Z, or any lettertogether with a single digit:

AO, A1,... A9BO, B1,... B9

ZO, Z1, . . . Z9

So to evaluate the expression given earlier,

(3.14+24492 -4.6x 106

the program given here would also be suitable:

10 LET A=3.14+ 2*1/220 LET B = 9*9 - 4.6E630 LET C = A T 4/B40 PRINT C50 END

THE PRINT AND INPUT STATEMENTSThe PRINT statement has already been met, but isnow considered in greater detail. One or more itemscan be printed by a single PRINT statement. The line100 PRINT 2/3, X, 2.3 will type out the value ofthat is 0.666667, followed by the numerical value of Xfollowed by 2.3. Since the items in the PRINTstatement are separated by commas, they will betyped out in columns across the page. Alternatively,if they are separated by semi -colons then they will besqueezed together. Anything which appearsbetween quotation marks in the PRINT statementwill be printed out exactly as it appears in theprogram. Plain numbers and arithmetic expressionsare allowed in the PRINT statement. A typical line ina program would be:

50 PRINT "THE ANSWERS ARE: ", X; Y,39/3; Z

This will cause the following to be typed when X =0.333333, Y = 100, Z = 6

THE ANSWERS ARE: 0.333333 100 13 6

The effect of 60 PRINT is to advance one line ofpaper. If a PRINT statement ends with a semi -colonthen the printing head will remain in its last positionuntil another PRINT statement is encountered.

The INPUT statement is a method of assigningvalues to variables from the keyboard as the programis running. When an INPUT statement is reachedduring the execution of a program, the program willstop and continue only after a number has beentyped in at the keyboard. The variable mentioned inthe INPUT statement will then have the value whichwas typed in. The input of a number is prompted by aquestion mark being typed.

So then, if it is wished to read in a value for Xduring the execution of a program, this is done bythe statement:

10 INPUT X

More than one variable may be assigned a value by asingle INPUT statement:

20 INPUT U, V

When this line is reached in a program, then it willstop and continue only after two numbers have beenentered. U will have the value of the first one and Vthe value of the second one.

On many systems it is possible to change the

prompt offered by the INPUT command by placingthe required replacement in quotation marks in thestatement thus:

10 INPUT "TYPE IN THE VALUES OF X ANDY: ", X, Y

Note that the comma is required after the lastquotation mark.

Example: This program will read in two numbers andtype out their average.

10 INPUT A, B20 PRINT (A+B)/ 230 END

To execute the program, the command RUN isgiven. At line 10 values for A and B are required. Therequest for these values is indicated by a prompting'7'. The values must now be entered with a commabetween them, followed by 'carriage return'.

RUN

?4,65

END AT 0030

Now to change the program some of the lines areretyped:

10 INPUT "TYPE IN THE TWO NUMBERS: ",B

20 PRINT "AVERAGE IS: "; (A+B)/ 2

If the command LIST is given again, a list of theamended program is obtained. A run of the newprogram will produce this:

RUN

TYPE IN THE TWO NUMBERS: -4.2,3.9AVERAGE IS: -0.15

END AT 0030

THE GOTO STATEMENTA program is normally executed in order of the linenumbers. The GOTO instruction, however, causesthis order to be changed. For example

50 GOTO 10

This statement will cause the program to jump to line10 rather than carry on to the next line after 50. Anyexisting line number can be used in the GOTOstatement.

Example: This program will type out the naturalnumbers, together with their squares andsquare roots.

10 PRINT "NUMBER", "SQUARE","SQUARE ROOT"20 PRINT"30 LET N = 140 PRINT N, N*N, SQR (N)50 LET N = N + 160 GOTO 4070 END

Work through this program and you will see that itcontains an infinite loop lit will keep running forever. The only way of stopping this program is topress the 'escape' key during execution. Naturally,

22


this is considered bad programming technique, but ithas served a useful demonstration purpose here.

CONDITIONAL BRANCHINGTo GOTO statement provided us with unconditionalbranching, but it is possible to ask a question beforebranching, and branch only if the answer to thatquestion is yes. It would be useful in the exampleabove to ask at line 45 the question 'Is N equal to 10?'- this would avoid the infinite loop. If the answer tothis question is yes, then the program will stop.Otherwise the program continues to run as thoughthe question had not been asked. The question isasked in BASIC by the statement.

45 IF N = 10 THEN STOP

The flow -chart for the program now looks like this:

START

PRINTHEADING

LET N = 10

PRINT N, N*N, SQR(N)

YES

NO V

LET N = N + 1

rSTOP

The program is:10 PRINT "NUMBER", "SQUARE","SQUARE ROOT"20 PRINT "30 LET N = 140 PRINT N, N*N, SQR(N)45 IF N = 10 THEN STOP50 LET N = N + 160 GOTO 4070 END

In general, a question may be asked containingany of the relational operators:

> = or = >< = or = <>< or <>

equal togreater thanless thangreater than or equal toless than or equal tonot equal to

A test of this sort can result in any of the BASICstatements being executed,

e.g. 20 IF A> = B/C THEN PRINT A

This will result in A being printed if and only if A isgreater than or equal to (B divided by C).

110 IF A*SINIT)<=10 THEN GOTO 210

If the result of this test is true, then the program willjump to line 210, otherwise execution will continue asnormal.

THE REM STATEMENTWhen writing a long and complicated program, it isoften useful to have explanations along the wayactually written as part of the program. This is donein BASIC by the REM statement. If a line begins withthe letter REM (after the line number), then that linewill be ignored when the program is executed.

e.g. 10 REM THIS IS A REMARK WHICHWILL NOT INTERFERE WITH THEPROGRAM

20 PRINT "THE SAME AS USUAL"30 END

FROM HERE?Although only a few of the facilities available inBASIC have been described here, the reader shouldnow be able to write programs to solve even quitecomplicated problems.

BASIC was designed for the user in the hope thatthe computer should be used as a tool to findsolutions to what might otherwise be insolubleproblems or at least problems which would involvelarge amounts of time when tackled by hand. But theuser often becomes interested in computer pro-gramming for its own sake, finding it intellectuallysatisfying to have written a concise and effective pro-gram. By writing simple programs in BASIC one canbegin to appreciate the pleasure which can beobtained from computer programming.

23


Typing Without TearsTo many people, typing is an integral part of theirjob, but to others the typewriter is merely a means ofproducing reports, manuscripts and so forth withoutthe need for the recipient to decypher thehandwriting of the author. To the new breed ofcomputer hobbyists, typing is becoming more andmore important as more time is spent at theteletype/VDU. Until machine readable media arecheaply and widely available, the only way ofentering a program will be to type it manually, line byline. To type with one finger may be easy to learn andone soon becomes deceptively proficient so that onedoes not realise the advantages to be gained by'touch-typing'. A keyboard can be a very efficientinput device and it is sacrilege to negate this inherentefficiency by typing one fingered.

The routines presented here are not to be regardedas a substitute for proper teaching but for anyonewho has to teach themself, they offer rather greaterchances of success than typing repeatedly easilylearned exercises from textbooks.

At this point, the text was to have read 'the authorcannot touch-type and has despaired at severalattempts to do so.' However, since this article was firstdrafted, many moons ago, I have in fact made theeffort to learn and I can attest to the greatimprovement it has made to the satisfaction I derivefrom the writing of an article. Needless to say, thislearning has encroached upon the time allotted to thecommissioning of my home-brew 6800 system, butwhen I eventually get around to it, I shall implementthese programs to improve my speed and, moreimportantly, my accuracy.

24


C START

cueeENT '.="A

NEXT "a"

OUTPUTCURRENT.

KEYPR E55E

7

yes

NEXT= NEXT.

NEXT .44:1"

cukzem- NEXT

Figure 1. THE BASIC TYPING PRACTICE ROUTINE

of which is a derivative of the first, and also rathermore useful. However it is instructive to examine firstthe simple version and then to see how it can beextended. The practice in this case takes place fromall the letters of the alphabet, which may or may notbe a good idea for someone starting to learn to type,depending how well they respond to a challenge andwhether they merely throw up their hands in despair.A flowchart for this program is shown in figure 1.Where letters are shown in quotes, this means thatthe ASCII value for that letter is being referred to.The program variables whch are very important are'CURRENT' and 'NEXT'. These, as their namesimply, are the variables which are used to store thecharacters which are currently under considerationand the character following it respectively. Why it isnecessary to look ahead will become clear in duecourse. The computer outputs the current letter,initially an 'A' and then goes into an input loop until akey on the keyboard is pressed. On each passthrough the loop, the variable 'NEXT' is incrementedand the tested whether it has become greater thanASCII 2'. If so, it is reset to 'A' and the input loopcontinues as before. Since the time at which the userinterrupts this loop by pressing a key is essentiallyrandom, what has been done is to create a randomvariable, 'NEXT'. As was mentioned before, 'NEXT'is the variable containing the character for the next tobe considered so the result is that this character hasbeen randomised. This is of course very useful sincemost so-called random number generators areessentially cyclic in nature and could present the userwith easily recognised patterns. This completes allthe hard work and all that has to be done now is tocopy 'NEXT' into 'CURRENT' and to jump back tothe beginning of the program.

This program has obvious appeal in its simplicitybut it gives no indication of whether the user hastyped the correct character or not and it will onlywork with the whole alphabet, or by changing thelimits (in this case 'A' Et 'Z') in the input loop to linearchunks of the alphabet. A revised version of theprogram is shown in figure 2. This embodies the twoimprovements implied above, namely a count of thenumber of correct and incorrect characters typed(positive reinforcement) and it allows the user tochoose the specific letters of the alphabet which hedesires to practice. The basic principle is exactly thesame, that is that the length of time the user takes torespond to the prompt is used to select the nextcharacter used. However, instead of 'NEXT'containing an ASCII character, it is used as a pointerto a table which contains the wanted characters.Also, when the key on the keyboard is pressed, thisis examined to determine whether it is correct or notand either 'PASS' or 'FAIL' is incrementedaccordingly. In the previous version the only way theuser could escape from the program was by somephysical reset button on the processor but now theprogram looks for a carriage return character andwhen it detects one, jumps out of the main programand prints the number of passes and fails.

TWMTL",,,

Figure 2. THE MORE ADVANCED ROUTINEWHICH HAS ADDED FACILITIES.

(*: INSERT HERE THE No. OF CHARACTERS IN THE TABLE).

To pass now to the rather more esoteric subject ofactually teaching oneself to type, the overridinglyimportant factor is to achieve a rhythm which willbecome much more important later as faster speedsare achieved. In days of yore the recommendedmethod used to be to type in time with a metronomebut these items are rather rare and very expensivetoday, especially when a 555 timer IC does the samejob at a fraction of the price. If one were to be reallykeen, the rhythm timing function could be writteninto the program itself but in this case, the hardware -software tradeoff is tilted firmly in the direction ofhardware.

25


4

Figure 3 represents a typewriter keyboard and thekeys outlined in bold lines are known as the homekeys. These are the keys where the four fingers ofeach hand come to rest when the hand in question isnot actually typing. The row of keys vertically (orobliquely) above and below each of the home keys isserviced by the finger associated with that home key,with the addition that the rows of keys between thetwo forefingers are also serviced by the forefingers.The space bar is of course operated by the thumband when a shifted character is required, the littlefinger of the hand not in use operates the shiftbutton.

This concludes the mechanics of the problem andall that remains now is a good deal of hard work onthe part of the learner. However, one or two other

applications of this type of program might be usefullydealt with at this point. A fairly obvious candidate forthis treatment is morse code. Both sending andreceiving can be taught, in the first case, thecomputer types (or displays a character) and thestudent has to echo it on a morse key attached to aninput port, and in the second case the computergenerates a morse code character and sends this to aloudspeaker and the student has to echo this on akeyboard, possibly improving his typing at the sametime. In each case, the length of time taken torespond to the stimulus can be utilised to decide thefollowing character. The same might also be appliedto the teaching of semaphore (if anyone uses it anymore), but the way in which this might be engineeredis left to the imagination of the reader.

" # s 7. & 9 C )

1 2 3 4 5 7 (E3 9 (75

Q R T I O

A S D c H K L< >

-Z XC VB NM 5 .

Figure 3. THE MAIN PORTION OF THE TYPEWRITER

KEYBOARD AND THE FINGERS USED ON THE

VARIOUS KEYS'

26

The mostsignificantevent aboutmicrocomputersin business andat home

Internationalexhibition,conference plusseminars andworkshops plusdebates.

1

CKSUNAL LOMPUTER WORLD

The 1978Do-it-yourselfComputer

SI

Show22-24 JuneWest CentreHotel, London

Please sendi me a fully detailedbrochure about

The 1978Do-it-yourself

Computer Show0 Dr/Mr/Mrsif

00

Initials

Name

Job title

Address

Return to :Online Conferences

Cleveland Road, UxbridgeEngland, UB8 2DD

Tel : (0895) 39262

27


The Mysteryof the LostComputer

Harry Harrison

I am beginning to feel that I once lived in anotherworld, one of those parallel worlds so familiar toreaders of science fiction. Perhaps because I writeSF I am beginning to believe in one of my own unrealuniverses. Why even my good friend, the goodDoctor Asimov, who knows everything abouteverything, doesn't seem to know about my world.In his book Science Past - Science Future hediscusses the computer revolution. He jumps in onemighty narrative bound from Aiken's Mark I, 1937-1944, to ENIAC in 1946. With nothing in between.Nothing? With quivering hands I picked up theglorious first issue of Personal Computer World andturned to "Past Procession" with its compact boxedhistory labelled COMPUTER CLOCK. The truthrevealed at last. But unhappily for me the truthoutlined there is the same one that Isaac Asimovrevealed. Yet I put it to you that both of these learnedgentlemen are wrong. There were computers ingeneral use between those dates; hard-working,practical computers that did their job remarkablywell.I should know, since I was a computer mechanicduring that period and serviced these machines.What both Asimov and Kewney have ignored for themoment is the fact that there are more ways ofrunning a computer than by electronics and digitalmeans. What about Mr Babbage? With that clue inmind I challenge all you computer fanatics out thereto figure out what kind of computer I am talkingabout. Cover the next lines with your hands andthink - no cheating! Have you thought? Good, theanswer is. . .

. a mechanical analogue computer of course.I have before me, stained and dog-eared, a UnitedStates War Department Hand Book on the operationand maintenance of The Sperry K-3 AutomaticComputing Sight. Something that computes is a

computer, right? And when I worked on these thingsmy technical MO was Power Operated Turret andComputing Gunsight Specialist, Skilled.Since I once saw one of these gunsights on sale afterthe war in a surplus shop, I hope I am breaking nolaws by quoting from the hand book. (Dated Sept.1943, I must add.)"The sight computes the range and automaticallycompensates for ballistics and prediction. . .datainput is transmitted to the computingmechanism. . .After the complete solution, the lineof sight is offset the proper amounts from the line of

The enemy targeted neatly.

The proof that cleanliness is a major virtue, even in computing.

If the computerbe washed withalcohol, benzine,white gasoline,

or carbontetrachloride.

excessively dirty it may

28


bore, in lateral and vertical planes, to compensate inall deflections."That last bit, translated from the militarytechnicalese, means the sight points one way, theguns the other. Like leading a bird with a shotgun.When the trigger is pulled the bullet rushes off to thespot where the plane is going to be, not where it is atthe moment of firing as with a stationary target. (Andlet us not be too sure of that stationary bit, either.The Paris gun, with which the Germans shelled thecity in World War I, fired a shell that travelled a higharc 92 miles long and took 3 minutes to reach itstarget. In order to hit the city the rotation of the Earthhad to be considered in the computations.)

Heath -Robinson CollectionThe only constant the K-3 worked with was themuzzle velocity of the guns and the resultant ballisticorbit. All of the other variables were either fed in bythe navigator, the gunner or the instruments; altitude(therefore air pressure), wind velocity, distance totarget, relative speed of target, and more. Thesevariables were inserted by the strangest Heath -Robinson collection of gadgetry ever conceived byman. Gear trains, take-up springs, revolving plateswith offset ball cages and - pride of the pack -three dimensional contour cams with cam followers.These were beautiful things, machined out of solidbrass, and could be rotated about a central axis, ormoved back and forth along that axis. They were themost educational device in the world because youcould see the analogue actually at work.But our computer problems weren't the ones youhave today. No glitches or static electricity problemsto bother us. Backlash in the geartrains was what wehad to worry about! Not that we didn't haveelectricity in the K-3 - it was not powered by steamand leather belts as has been hinted - why, we had

a constant speed 2000 RPM motor. And a light bulbfor the reticle sight.The computer worked. These same principles wereapplied in that other analogue computer, the NordenBomb Sight. (Someday I must tell you about thecomplicated device it used to release the bombs.)The B -17's and B -24's that carried power turrets withcomputing gunsights flew in stepped formation soany attacking pursuit plane would have to face thefire from a number of guns. The German fighters, forall their strengths, were very unstable gun platforms.They had open sights and their wing guns werezeroed at a point 200 yards in front of the plane; thatis, the cone of fire from the guns met at this point.The maximum range where they might hit what theyaimed at.The K-3 was accurate from 1000 yards on in.So the aerial gunner had theoretically 800 yards tofire without being fired back at. Thirty shots a secondfrom his two calibre .50 guns firing armor piercingammunition. The sights worked and they did the job- we won the war, didn't we?So don't let anyone knock mechanical analoguecomputers. If the airborne argument doesn't suityou, why, just think of the Senior Service. All thatbeautifully accurate fire and great ranges, manytimes over the horizon of the guns firing, wascontrolled by "fruit machines."Read "Mechanical computer" for that last and you'llhave a good idea of where the computer action wasin the forties. It wasn't until 1945 that I saw my firstelectronic computer. This was the Central FireControl system on the B-29, the plane that droppedthe atomic bomb. All thermionic valves. I had no ideaof how it worked, I would have had to re -enlist tofind out and civvy street's lure was too strong. I mustfind out some day - because I notice that thehistories of computers don't mention this one either.

Then oil ball carriages, discs, drumsand felt wipers with Univis No. 48

29


EXPLOIT ORCUtlIVATE

COMMENT ON THE PERSONAL COMPUTER BUSINESS

Guy KewneyThe sad truth of the matter is that there is still noretail computer market in this country.

Retailers in most businesses expect to makebetween 35% and 45% profit on what they buy forresale - depending on the perishability of the goods.

Many of the new businesses just struggling tomake ends meet by selling cheap computers in thiecountry are trying to make well over 100%.

If you don't believe this, just compare prices ofAmerican kits and systems as advertised in Byte, DrDobbs, Kilobaud and so on, with any price you maybe quoted for the same kit from a 'retailer' in thiscountry. Typically, the sterling price is nearly thesame number as the dollar price - and remember,the dollar price already includes at least 25% profit,and importers here typically get a substantialdiscount. In one case I know of, the UK distributorgets a bigger discount than US agents, to covercosts of importing.

This is not retailing - or if it is, it is dishonestretailing - but the computer business in all its glory.

If you program computers yourself, or designcircuits, you may not realise what 'all its glory'actually means. I will give you one example of thesort of price structure you get in the computerbusiness.

Three years ago, working on the now -defunctjournal Dataweek, I was shown a big disc drive beingenhanced by a company which built systems arounda famous make of minicomputers. With a specialinterface, this company sold the disc drive at around£8,000 - they built the interface themselves, andthey bought the drive and the controller bog -standard from its supplier, for £3,000 or £4,000.

This drive and controller probably cost a maximumof £1,500 from the factory, estimated my friendlysystems builder company. And, did I know, the samedevice formed the disc drive sold by company X?What did I think X sold it for?

X sold it for around £28,000.I didn't believe it either, so I checked. First I

checked with X themselves, who said yes, but I hadto realise that the sale price also covered . . . andwhen they finished their list of excuses and I askedthem what the cost price was, they refused tocomment.

So I checked with the supplier and they admittedthat the drive was basically the same machine as theone they sold, in small quantitites, for under £4,000.

But they said, although they sold it in quite largequantities to X, it wasn't quite standard - it had adifferent board of electronics components in it, tothe company X's specification. No, they weren'tprepared to say what it cost after the bulk discountwas subtracted and the special board was added.

My editor decided not to print this story - notbecause he didn't believe the figures, or thoughtthere was some extra factor which ought to be addedon the costs - but because he felt it would be unfairto publish the company X's name on this sort ofmargin if I couldn't also show that another company

Y did the same sort of thing - which he himselfdidn't doubt, by the way.

You don't have to look far for further examples ofan industry which needs this sort of profit margin tocover its gross inefficiences of design and thebeyond the frontiers of technology, both hardwareand soft. Why does a video display with fewer com-ponents cost four times the price of a colourtelevision? Why does a matrix -needle printer withvery few moving parts cost twice an ordinary motor -driven typewriter's price? Whatever your explana-tion, it better not be 'they cost more to make', Theydon't. They cost less.

Well, that's the computer business, and it works,on its own bizarre level.

It won't do for the computer retail business.It's all very well for those who are just setting up

their own businesses in this area, without a lot ofcapital behind them, selling ones and twos to startwith, and having to cover the mortgage on a largefamily house and everything that goes with it. Theycan say that, selling to polytechnics, the HomeOffice, and wealthy central heating distributors,these margins are acceptable to the customer, andthat without them, the business wouldn't get off theground - but I wonder.

I tend to agree with Nen Norman (she won't tellyou what Nen is short for and I promised not to) ofComputer Centre, recently opened.

This lady started out her and her partner's businesswith an 8K RAM board for the Altair -type S100 bus,costing £85 in kit, £95 assembled - static, notdynamic. I announced my pleasure at finding areasonably priced memory board, and incautiouslytold readers of my Personal Computing column thatthe Centre must be making only about £10 on it.

It just goes to show how you can get used toexcess profit margins, I suppose, because the inkwas still flowing over the pages when Nen's partnerrang me up to explain that he was not making only

- nothing like it. 'A comfortable profit' wasthe phrase used, and we left it there.

But the Centre people then expressed themselvesforcibly on the subject of other operators in thecheap computer business. The quote I have down inmy notebook goes:

'There are far too many people guilty of sup-pressing the development of the amateur and privatecomputer business, by overpricing.'

I agree.The business is not going to take off until these

operators realise they are in commerce, not inindustry, selling to consumers, not clients - and,more important, that there is more money in it.

Perhaps I better say that again, for the slow andhard of hearing. There is more money in a bigconsumer market than in a small buying clique.

This isn't just a question of the economies of scale.It's a question of publishing, and software. On theface of it, there may be nothing to choose between

30


selling 50 computers and making £400 profit on each,or selling 500 and making £40 on each, except thatthe 500 take a lot more paperwork to shift.

But there is a very important difference. Each ofthe users of the 500 can pick 499 other brains andsoftware libraries, and operate quite a lot quicker inachieving whatever they bought their system toachieve.

And after two years, when both sets realise thatthey can buy 100 times the computer power for halfthe price, and both the suppliers have to design MarkII or lose their customers, the 50 -man will have only50 customers already hooked on his software, whilethe 500 -man will have a nice -looking user base -and probably a user base who are a lot moreenthusiastic about the product. For one thing, they'llhave more peripherals, more software, more generalideas . . and for another, they can look forward to acheaper replacement machine.

While the 50 -set will have come to think of theirsupplier as something of a rip-off merchant.

The high -cost people are not rip-off merchants, ofcourse. They are merely cautious - overcautious inmy view - and don't yet really believe that the man -in -the -street wants to spend £300 on a computer.

They'll learn.

BookReviewMichael JamesTHE BEST OF BYTE VOLUME 1, edited by D.H. Ahl & C.T.Helmers, 1977,- 376 pages. (Creative Computing Press, paper123.00)

'Best of Byte Volume One' was published in March '77 andconsists of a selection from Byte magazine from Sept '75 to Dec'76. Thus the usual glimpse of the future, expected by anyoneused to reading up-to-date US publications, is sometimes missing.However, as a view of what was going on in the USA and hencewhat is (just about) going on here it is excellent. It is split intoseven main sections, Opinion, Theory and Technology, ComputerKits, Hardware, Software, Applications and Resources, so readershave no excuse for not finding something of interest.

The longest section, with some 148 pages, is devoted toHardware. Most of the articles do not give complete con-structional details (i.e. printed circuit layouts are the exceptionrather than the rule) but the designs are generally well explainedand should prove easy to interpret. My own vote for the mostinteresting item goes to the 'fast cassette interface'; working at1100 baud, it would save a lot of time when used in place of a 300baud Kansas City standard system.

The Software section at 114 pages is nearly as long, and if hard-ware is not your speciality then this, along with the remainingsections, should provide a good introduction to the non -solderingpart of computing. If you would like to know how an assemblerworks, or how to calculate your biorhythm chart, or design an on-line debugger, then 'Best of Byte' will tell you. There are fewcomplete programs listed but plenty of ideas and 'how to do it'notes. The description of how to compile algebraic expressions isnot quite complete, but very easy to follow and succeeds inrevealing a lot of the computer science inside a compiler.

The Computer Kit reviews are useful but are descriptions ratherthan evaluations, and a number of important recent machines arenot included. The following are covered: the Altair 8800, SWTPC6800, Altair 680, KIM, Z80, and the LSI 11. It is an interestingreflection of the differences between UK and US home computingthat the LSI 11 at about £1000 is included as a computer kit! Themost disappointing section is the one dealing with applications.With only two articles and nine pages it is, compared with theother sections, severely undemourished and leaves one thinking'what are home computers being used for?'

At £8.00 'Best of Byte' is good value, and its large A4 sizedpages hold a wealth of knowledge, opinion and ideas. It can berecommended to all but the absolute beginner.ends tape 61

Y=oCOS(VX/b).a.81, b=42.86

IF YOUR COMPUTERCAN'T SMILE <ORPLOT) IT PROBABLYHAS MERLIN DEFICIENCYANEMIA.

(EQUATION PLOTTED WITH BASIC)

NOW THE FAMOUS MERLIN WITHSUPERDENSE IS AVAILABLE IN THE UK -and with improved graphics resolution and 25%more lines of text!On European standard TV sets the bestASCII /graphics boards on the market for theS100 bus give 25 lines of ASCII or 320horizontal x 250 vertical points of DMA bitmapped graphics.

If you have been considering an ASCII/graphicsinterface for your system, or if you hadn't beenconsidering one because you thought goodgraphics was too expensive, consider theMERLIN with SUPERDENSE add-on. It's thebest there is and its at an affordable price . .

Kit, only £275

CONVERT A STANDARD SELECTRICTYPEWRITER TO AN RS232 TERMINALThe Sharp 8- Associates kit to convert an officeSelectric is now available in the UK. Full kit withreader/punch port and RS232 interface,conversion instructions and manual, only£540.00

For full details of these and many options landalso our other personal computer equipmentsuch as Assembled 8K memory boards for £120)

Write or Phone

J Et A COMPUTERS15 FLEETWOOD GARDENSMARKET HARBOROUGHLEICESTERSHIRE0858 7620

31


COMMODORI 'S APPROACH TO COMPUTINGKit Spencer

THE EARLY PET STORYThe story of PET begins with the development of the 6500 series of microprocessors by MOS

Technology which then became the heart of the KIM I microcomputer system. Less than eighteenmonths ago Commodore acquired MOS Technology to give us our own in-house chip processing anddesign. Since then KIM I has become the best selling microcomputer system in the USA. (It is nowavailable in the UK for £149 plus VAT).

PET effectively graduated out of the KIM I - the whole project taking less than a year from the firstdevelopment "go ahead" given by Commodore after we realised the vast potential such a systemcould have and the fact tht we now had the necessary in-house technology, production andmarketing capability for the fascinating PET project.

What Was The Design Philosophy?Firstly, let me make it clear that this is not a

technical paper but an insight into our overallphilosophies. Several main concepts were used indesigning our PET. These were that it should beAFFORDABLE, CAPABLE, EXPENDABLE, UNDER-STANDABLE, USABLE and SERVICEABLE. Let meexplain each of these concepts more fully.

CapableAlthough 'popular' media sometimes talk of PET in

terms of game playing, I should make it very clearstraight away that PET design parameters were notthis but that PET should be just as capable as atraditional computer that up till now would have costthousands of pounds. The PET was firstly designedto a demanding specification and then we set aboutdeciding how we would get the price right for whatwe wanted to achieve.

AffordableWe felt that if we could offer a full computer

capability at under f1,000 the market for computerswould change from one of 'major company boardroom decisions on capital equipment' to a level ofpersonal and departmental decisions that wouldopen up dramatically increased market applicationsand potential. If this was so we could write offdevelopment costs over many more units and applysome of our mass production and marketingexpertise that would make this an economicproposition for us as well as our customers -response so far appears to justify fully this decision,thank goodness!

ExpandableWe designed our PET computer so that it could be

32


upgraded as the heart of an extremely powerfulsystem. For this reason we built in an IEEE -488output (often known as the H.P-I B) which allows formultiple intelligent peripherals - we have underdevelopment our own printer and floppy disc.

The 8K Ram of the machine is designed to beexpandable in 8K increments up to 32K and thesewill in the future be offered as optional extras. Thereis also a parallel user port and a second cassetteinterface (the second cassette is another optionalextra. I 0 is managed by built in operating systemsoftware.

We believe all this makes our computer systemvery expandable.

UnderstandableWe have tried to take some of the mystique out of

computers in order to broaden people'sunderstanding, appreciation and potential use. Forthis reason we chose an expanded, fast,comprehensive and powerful 8K Basic which hasbeen built into the machine (part of our 14K Rom) asthis is one of the easiest computer languages tolearn, understand and use. There is also machinelanguage accessibility for the professionals.

Our PET computer has been designed as a fullyintegrated and independently operable unit that canbe used from the day it is purchased by simplyplugging into a standard mains supply. For thisreason a 9" video display, built in cassette andkeyboard were designed into PET - PET can beused as an excellent teaching machine and, in itsdirect mode, is an excellent programmablecalculator. (There is even a built in clock for timingapplications - settable directly from the Basic via asimple keyboard entry.)

PersonalThis is perhaps one of the slightly misrepresented

aspects of our PET computer. It does apply to the

price of the unit, to the fact that it is portable and tothe fact that it can be operated even without any ofthe expansion capabilities being added - it certainlydoes not mean that it cannot be used other than as apersonal unit. We know the unique graphicscapability makes it a fascinating 'personal computer'for the hobbyist programmer but that's only oneaspect of PET.

ServiceableWhen you open up a PET to look inside you will

immediately realise it has been designed for ease ofservicing and economic mass production techniques.

The PET consists of only three p.c. boards - themain board which can be snapped out and easilyreplaced in the event of any defect and the other twoboards which are for the CRT circuit and the cassetterecorder (specially modified), both of which areremovable and replaceable for ease of servicing -and assembly, of course!

The other components are a power supply, CRTand a keyboard. The PET has been designed for theperson who wants to use its capabilities and not forthe man who wants to poke his soldering iron around- at least inside the machine.

What Is The Future For Our Pet Computer?As an International Corporation we have already

made major investments in this market and ourefforts will be continuing in three areas. peripherals,software and marketing. All of which are going to beextremely important.

What About Peripherals?There are no specific details yet (I don't want to be

misquoted) except to say that our first plans, alreadywell under development, include a second cassettedeck, a printer, memory expansion and floppy disc.But remember, if you don't like our units or can'twait there are others compatible with PET. For those

33


of you who can wait, the prices are not likely to beout of proportion to the very reasonable cost of thePET computer itself. We do believe in passing on thebenefits of mass production to keep us ahead.

What About Software?We recognise the importance of this area which isone of the reasons we have started our UK PETUsers Club already. We recognise the fact that someof the best software created for PET will probably notcome from us but from those of you who buy PETsand develop really useful programs for your ownneeds. We want users to share this with all ownersand are offering four free programs for everyprogram sent to us that we accept into the library.Coming with the computer will be a programmingmanual fully describing our excellent Basic inaddition to the general manual. The Users Club willalso be regularly putting out extra information on theareas users tell us they most want to know about.We don't intend to inflate the price of our computerby charging you initially for a lot of very expensive,specialised software we have developed and youmay well not want.

In addition to Basic software, we offer a verycomprehensive programming manual on themachine language of our own MOS 6502microprocessor, which is at the heart of PET, soprogramming can be done in Basic or Machinelanguage.

Incidentally, not everyone is aware that lower caseletters are available within PET via POKE commandsthat switch the upper case functions between thisand the graphic functions actually printed on thekeys.

What About Marketing?One of the key elements in achieving the

breakthrough we have in the price specification ratiois to take a different approach to the market frommany traditional computer companies. We do notintend to employ a several hundred strong sales forceto wine and dine would-be customers and, ifsuccessful, get a commission quite possibly morethan the price of our own computer for what inspecification is a similar unit. We believe that it isbetter to pass this saving on to the customer and, ifhe is as sensible as we happen to believe he is, thePET itself and the owners are going to be the bestsalesmen we can have - again, initial reactionseems to prove we are right.

However, we do recognise the importance ofcertain types of marketing which is why we have setup a completely new division of Commodore here inthe UK (Commodore Systems). This will handle PET,KIM and allied products under development.Commodore Systems will be a small team ofspecialists designed to back up and develop themarket for an excellent product and not a big salesteam to push another 'me too' product.

As production and our organisation grows, weintend expanding to a local dealer network to givegrassroots support - these will be carefully selected.Meanwhile we are already setting up new offices anda "PET Computer Centre" in Central London (opensometime in April at 360 Euston Road) which wehope users and others interested will visit to evaluatePET, exchange ideas and generally use as a

'Personal Computer Centre'. We'll let PersonalComputer World know just as soon as we have amoving date fixed.

APPLE 11 MICRO COMPUTER

Simplythe best!

Apple II comes in a portable, 10 lbs. zip caseand is the ultimate personal computer,

with powerful colour graphics, extended Basic,8 I/O ports,1000 cps. display, 8K ROM,

up to 48K RAM, 350ns memory access,mini assembler and much more!From £995, inclusive of cables,

manuals, software. Maintenance arranged.

Der8onalComputer

Limitec.18-19, Fish Street Hill, London EC3R 6BY.

Tel; 01-6231434. Telex 88264.

SC/MPMICROCOMPUTERSYSTEM f99.50 +VAT

If you're in engineering, software or both, are a student orhobbyist, or simply looking for a high-powereddevelopment system for National's SC/MP at a realisticprice, read on!

HARDWARE 7 segment display for every register (32 digits!) 20 key keypad Completely transparent operating system 'AK RAM, expands to 64K Buffered system bus interfaces with cassette/TV/TTY

peripherals etc. via plug-in motherboard. (LCDS/ETI buscompatible)

Single step, slow step facility

SOFTWARE Operating system allows easy modification of registers

and memory, move and copy data blocks and much more Set breakpoint register will allow traps on any location,

including data and ROM Software breakpoint

For full details of the system and supporting products anddocumentation write to-

I.E.S. LTD35 Purleigh Road, Corsham, Wilts

or phone

(0249) 713515/712317

34

law

Guy Kewney editor of the only weekly Personal Computing newscolumn in Europe at home with computer and his daughter, Lucy.

"Every week in Computing I assemble andedit a page of news about PersonalComputing."Frankly, it isn't meant to be the accumu-lated wisdom of the ages -- though I do try toget quite a bit of depth to my stories."What it is meant to be is up-to-date - oreven a bit ahead of time."Perhaps somebody has already written thatbit of extra software you are thinking ofstarting, or will have announced a processortwice as powerful as the one you are writinga cheque for, at half the cost, or they havesolved that memory expansion problem; orthey went bust last Thursday.

To Julie Squire, ComputingRoom D11 Haymarket Publishing Limited76 Dean St. London WI A 1BU

Please send me free of chargea copy of Computing every weekI am not a member of theBritish Computer Society.

NAME

ADDRESS

L _

By the time

you'veread this.....

"At the speed the personal computer worldis moving, you certainly need the soliddetails, construction hints, and software youwill find in Personal Computer World. Butyou also need a good weekly update."Not only do I run more than one system ofmy own, but I also make sure to talk to thepeople who matter in the businesssuppliers, new companies, users."By the time you read this, a lot more willhave happened that you need to know about."Don't miss your weekly update. HHere's some of the news you may havemissed, or got too late if you haven't beenreading the Personal Computing page...

mothut

into teething. troublesseries of articles m a errors

I hen hirlipoinck his outSystern 68 computer and board.' 7aidth'Ifi'll:ti'yr.

at many complainantsfor

welmlwo°rYrthba9

in a UK do.it 'most of which hare beene wrong, mitgounecorrected in 61'

note of yuenily,etle-S We

When started this

beirg overambitious.

oil:einernfiTylearnibeboa"rd'k'irit rolirlg.3,Nit vac at Sintromham the compmer

centre. A hg hall tar me Ymiblot to Hants. It tits an Icrmpwm' mvMtthe dery

Thu must hewer shop al0 a

age' ego... 01 March.gAM

En;.tad, which

r:wain:7e

mamma,,t rang'

rt with Me rangemade by South

8, -

7221S5

-1-y system runs

an,

the selotiusila MOO miff. the mire. oh, sill be MiCdprOCCS. the Sternum AficroMop. nest

This o hind ause door to the company Immosomething 61 ow In the in ending.somicoodusior ommess. He told me' 'The man se

Sintreni is also oilman will lime to he enthosystems distribution agent Mta omit: and an onto:tumourIMO- hut, as Daltel I mints - the .cwt of man silos antsout. the company .as supply. tort hu com small bus,ing Molorolabased mui newsta but doesn't hme the

helong Odom it took on capital hoc., 'Intel transfuse Poihably. said Wilco

II is now

Mete,prubahl

pmmgcondo

SW

S.

uithcoontrwill

'computing is sent free on request to people involved in the computer field in the U K Others moy receive Computing on subscription for 1.15 a yearinclusive of oil postage charges


AVERY SMALL

BUSINESS

SYSTEM?Martin Collis

In the past, most software houses in the U.K. havebeen unwilling to invest in hardware, the usualreason given being that installing a system wouldaffect their independence. In practice in-housesystems must greatly increase productivity and, ashardware prices continue to fall, it will soon becomenormal for most software houses to own a

considerable amount of hardware. As we are a verysmall software house, we have started by installing avery small system, an SWTP 6800 with a teletype ©and twin mini floppy disks.

Our reasons for installing a system were: -1. To see if we could put a computer kit together

and make it work - we have no hardwareexperience.

2. To see if a very small system could be used forcommercial applications without the lowhardware price being swamped by softwareand maintenance costs.

3. To have an in-house system available fortraining.

4. To get some experience, and hopefully somework, of micro software development.

We decided to base the system on a teletype © as, ifall else failed, we would at least be left with aterminal which we could use for punching papertape, time sharing development, etc. We were luckyto be able to buy a nearly new Data Dynamics ASR33 with a V24 interface.

At the time when we were looking for a system,which was last summer, the only possibilitieswere: -

1. The SWTP 6800 system from ComputerWorkshop

2. Three "S100" systems from Compelec, theAltair, Cromemco, and Polymorphic

We chose the SWTP system because: -1. It was the cheapest.2. There were (and still are) plans to manufacture

it in the U.K.3. The Motorola 6800 seemed to be more widely

used in the U.K. than the Intel 8080 on whichthe other systems were based.

The SystemWe ordered a basic system (processor, 4K ofmemory and one interface, an additional 12K ofmemory, a second interface and a cassette tapecontroller.

With the exception of an 8K memory board, andthe cassette controller, the kits were delivered aboutten weeks later. We bought some basic equipment, alow wattage soldering iron, wire strippers,multimeter, etc. and set to work.

With two of us working mainly in our spare time,the processor took about two weeks to assemble.The instructions were reasonably clear and oursoldering certainly improved with practice. It is quiteobvious from looking at the system, which parts weassembled first!

Testing Et DebuggingAs the SWTP 6800 includes a monitor in ROM and256 words of RAM on the processor board, the firstsystem tests can be performed without any memory.We assembled the interface board, connected theteletype and switched on. At this point the systemshould respond with "*" on the terminal. We got noresponse. We pressed RESET a number of times,this should have given "*" each time, still nothing.By poking around with the multimeter we establishedthat RESET was not working. I also shorted +12Vand -12V supplies on the mother board which blewthe power supply (but not the fuse!).

We took the system to Computer Workshop, whofixed the power supply and then established that theRAM chip was duff. They replaced it and the systemthen worked. This took about 45 minutes, was doneas soon as we arrived, and was free of charge.

We went home and assembled the two 4Kmemory boards. The diagnostics showed up faultson both of these which we were unable to fix. So wetook them back to Computer Workshop. This timethey were busier than when we went with theprocessor and we left them to be fixed. The faults, a

36


duff chip on one board and a solder bridge on theother, were again fixed free of charge, and theboards were ready within three days.

Now we could run the 4K BASIC, very basic, butthere was still no sign of the 8K memory board or thecassette interface. This was very disappointing asfour months had now elapsed since we had placedour order.

In a fit of extravagance we cancelled the cassetteinterface and bought twin mini floppy disks andcontroller as these were available from stock. The 8Kmemory board arrived soon after. We set to workagain and the disks and the 8K board both workedfirst time. This I suspect was luck but it might havebeen because we were now more experienced.

SoftwareWe now turned our attention to the software. Thefirst release of the mini floppy operating systemincluded BASIC but not data files. However as a newrelease of the operating system was promised soon(before Christmas!) we concentrated on getting avariety of games on the system. The BASIC is quitegood although it seems to be rather slow and wedidn't find any bugs. The system is kept in a roomwhich can get very cold and this seems to give rise toquite a lot of disk errors.

The new operating system finally arrived at the endof January. To our dismay it proved to be not version2 of the original system but version 0.9 (pre-release?)of a new system. The documentation was handwritten and, more dismay, we had to re -locate 4K ofmemory to 28K -32K addresses to run the system.

This meant that only 12K was available for BASICand the application program and we would have tobuy more memory to run reasonable sized programs

(shades of IBM). The final blow was that althoughdata files were supported they could only be read orwritten sequentially. We are currently waiting for thenext release of the operating system - we have alsofound a number of bugs in 0.9. We are also lookingat an alternative operating system which does offerrandom and sequential files.

To summarise our experience to date: -a) The hardware does seem to be robust and reliable.b) The SWTP system software still has some way to

go before it could be used as the basis for a smallcommercial system.

c) If you can go and see them Computer Workshopare very helpful in fixing hardware faults. Theirsoftware support is less good.Once we get a good basic operating system our

first step will be to develop a set of standard pro-grams for setting up, maintaining and printing datafiles, for generating indices and for selecting andprocessing records via the indices. This set of pro-grams will then provide the basis for a number ofstandard applications such as invoicing, word pro-cessing, stock control etc.

How much?A final word on costs. The processor, twin diskdrives and 24K of memory cost about £1,700 in bitform. A vdu and 30 cps printer would cost about£1,500 giving a total hardware price of £3,200. Ouraim is to provide a basic software package which theuser can extend to meet his requirements. It isdifficult to price this at the moment but it is unlikelyto be less than £800, giving a total price of £4,000.Using full size floppy disks would increase this byabout £900.

IN A FORTHCOMING ISSUEPCW REVIEWS THE TANDY TRS-80

*If what happened so far is anything to go by, customerrelations are high on Tandy's list of priorities. We have beenoffered a machine for review, for a reasonable length of time.

Barney Price of Tandy took delivery of the first two hundredTRS-80s in March. The TRS-80 is built around the Z80 micro-processor. In the photo above it's seen "in action". Tandy isoffering or is developing pre-recorded cassette programsfor a wide variety of applications. The TRS-80 microcomputerwith regulated power supply, less video display and datacassette recorder, is priced at £499.

And . . . just some articles in THE NEXT ISSUE*Interfacing a Westrex punch to your microcomputer.*Win the pools! At least, have fun with a "pools" program.*The Z80 instruction set.*Are you a microcomputer widow? More hints on how to program.*In praise of the PDP-11.

. . . It's the unique PCW mixture: articles for the beginner,hobbyist, recreationer, businessman and expert.(Please note that for reasons of space The Gates of Reasonand Buzzwords will appear in the next issue.)

THE APPLE II - Full PCW ReviewMike Sterland of Personal Computers Ltd., who believes that"Apple will become a major force in the Personal Computermarket", has generously loaned PCW a machine.

37


PCW CASE STUDYBoris Sedacca

The following case study should be of interestto the hobbyist but is not written specificallywith the hobbyist in mind. The intention here isto show that an average small businessproprietor can configure and develop a fairlysophisticated small business system around amicrocomputer.

Firstly, an outline of some parameters which wereset in the selection of this particular case; the over-riding constraint was that the total hardware costshould not exceed £10,000. Software was excludedfrom this pricing exercise because, apart from thefact that one would be hard-pressed to find a

workable system to include software under £10,000,the intention here is to show that one does not haveto be a "computer technocrat" to develop one's ownwith applications tailored to suit the business'sparticular needs.

What makes this a lot easier is the use of BASIC;an acronym for Beginner's All-purpose SymbolicInstruction Code. Basic's concise English -likecommands allow for quick mastery of programmingconcepts, and its interactive mode sends outprompts and error messages to the studentprogrammer at his terminal.

The main crticisms of BASIC are: slow run timeexecution through its interpretive stages (thecommands are basically sub -routines), the need forextensive documentation to append programs(COBOL, although much more verbose lends itself toautomatic documentation), and for moresophisticated versions of the language, high memoryoverhead.

A counter -argument to the first criticism is that thetime -related cost of developing a system usingassembly code programming is much higher than thesavings achieved. However, the economics are morecomplex if a program is to be copied a number oftimes as a standard package, where it wouldprobably be more cost effective to have a softwarehouse write an assembler version of the BASICprograms.

As for memory overhead, the reader will no doubthave heard of falling costs of memory due toMOS/LSI technology advances ad tedium.However, cost is not the only factor to beconsidered; most 8 -bit microprocessor CPUs havelimited addressing capability.

Not that BASIC is doomed to being an interpretivelanguage indefinitely; Jacquard for one has written aBASIC compiler to supplement its interpreter fordebugged programs to be converted to code. TheJacquard CPU is made up of a National Semi-conductor IMP -16 consisting of four 4 -bit slice chipsso the coded software is not cross-compilable.

The term "microcomputer" probably needs somequalification. The generally accepted definition forthis is that the CPU is made up of a single micro-processor chip. Another definition is in the use of LSI(Large -Scale Integration) as opposed to the MSI(Medium -Scale Integration) technology used in mostminicomputers.

Obviously there is a certain grey area as more andmore minicomputer manufacturers use LSI chips ontheir boards. As a further refinement to the definition

one could regard the microcomputer as a designwhere the chip is used to handle all operationscentrally on the classic von Neumann model, heavilydependent on the single processing resource. Thishas obvious implications in terms of instructionexecution speeds and is one of the many criticisms ofthe microcomputer in its present form.

This definition should not be taken too literallyhowever, as Fairchild's F8 would be automaticallydisqualified. At least three chips are required to makeup a CPU capable of addressing 64K Bytes ofmemory: the 3850 CPU chip proper, the 3851Program Storage Unit (PSU), and the 3852 or 3853memory interface.

Another limitation of the microcomputer in its usein commercial DP is its difficulty in interfacing to ahigh-speed DMA advice such as hard disk. Mostmicrocomputer systems in their present stage ofdevelopment handle their mass storage I/O througha multiplexer bus to floppy disk devices.

In addition to its relatively slow speed the floppydisc has a limited storage of 256KB or 312KB, theformer coming from IBM or IBM-compatible devices.IBM has recently announced dual density, double -sided floppies which brings the storage capacity upto just over 1MB per diskette. This allows forsubstantial expansion of file storage on their IBM5110 up to 4MB.

A microcomputer however should not be seen asbeing perpetually upgradeable to any number ofperipherals. If a business is going to require morethan 4MB (megabytes) of on-line storage and a fastprinter for a large volume of reports, it would bebetter off with a small business system offered by amainframe maufacturer or by a turnkey systemshouse.

Limitations aside it is possible to configure a smallbusiness system around a microcomputer, and anincreasing number of them are coming on themarket. Certain software houses are offeringstandard packages for these systems, written mainlyin assembler as they can afford the capital outlay ondebugging and development kit.

The following case study, then, shows how asmall business proprietor went about installing acomputer system at a price less than £1 0,000.

THE COMPANY: Computech Systems.THE BUSINESS: Computer staff recruitment

agency.The background: Computech was begun in 1966.

The original intention was to set up more than just arecruitment agency and to write software foraccounting machines and visible record computers.However the recruitment side became increasinglybusy and the other side was not developed as aresult.

Computech currently employs a staff of nine full-time employees and turns over in the region of£120,000p . a .

THE SYSTEM:EXTEL Microfile 32K processor ll£3,5002 PerSci floppy disc drives 312K 1-

30 cps EXTEL intelligent buffered terminal £1,950Upper & lower case VDU £750

16KB BASIC User licence fee £300

38


Operating utilitiesText editor Et elementary word processingFile handler

Microfile is built by a Californian Company, DataTerminals and Communications, and sold in thiscountry by EXTEL, the DTC U.K. agents.

The box itself houses the floppy drives andincludes slots for printed circuit cards which wouldinclude memory, the processor chip - Intel 8080A,the market leader and nearly the standard device foruse in such systems - and the electronics to controlthe interfacing of the VDU and the printer terminal.

The PerSci floppy disc offers faster data accessthan most, averaging 200 milliseconds to find andretrieve characters, where on most units the averagetends to be more than 300 milli -seconds percharacter.

Laurence Payne is the founder and proprietor ofComputech systems. He began business in 1959 in apartnership of Chartered Accountants butComputech was launched in 1966 as his firstcommercial venture.

"Although I did not find the time to develop anysoftware in those days I did buy and sell smallbusiness systems on an ad -hoc basis", he says."These were based on the Olivetti P203 and Adlermachines programmed for commercial applications.

"One could buy these machines extremely cheapsecond hand as the hardware value was low withoutsoftware. Software itself had no resale value.

"This does not apply with the new machinesprogrammed in BASIC as programming is muchsimplified here. Furthermore you can protect yourinvestment when selling the machine as it is made upof solid state electonics which do not deteriorate orbecome obsolete as quickly as the old hardwiredmachines."

As the agency business grew Computech sub-contracted its own accounting requirements to anaccounting machine bureau. The workloadcomprised mainly payroll and invoicing for a largeteam of temps.

L to R: Laurence Payne and Boris Sedacca

This system was found not to be satisfactory forvarious reasons. There was a slow turnround timeand lack of control over debtors. No records werereadily available at the office where they wereneeded. As the agency developed a policy decisionwas taken to discontinue hiring out temporary staffand the outside bureau service then ceased to becost effective. Finally, there were no interactivefacilities for registers of applicants. "The system weare using now is capable of both accounting andword processing applications.

"We developed our own text editing package tosupplement the Microfile's. The system now allowsfor the operator to start a new letter just by using thekeyboard and having commands appear on thescreen. These interactive facilities make this processcompletely self teaching.

'All the routines we've written are in BASIC

39


including multi-access for the connection of severalterminals. The standard version is only single userBASIC.

"We have a total of around a thousand applicantson our files which are held on diskette. Only aboutthree hundred are 'live' at any time and we can hold afairly large record of their particulars on one diskette.From these we can make parameter -driven enquiriesall under operator control.

"The link with word processing on these filesallows us to merge the files and send out a great dealof letters to clients and applicants. This morning wesent out about 130 letters most of them standard informat with variables such as names and addressesin the text. All we do is key in a few controlinstructions and the machine does the rest. Theoperator is free to carry on with other work while theletters are being typed out. Although the need for filesearching was the main reason for implementing thesystem, text processing has turned out to be a majorapplication."

Laurence Payne had started looking around forhardware since 1974. The concept for the systemhad existed long before but around that time therewere many minicomputers being announced whichmade it possible for a company with a total budget ofaround £30,000 to buy their own in-house computer.The problem here was that a system built around amini would have needed a great deal of effort interms of software and a software house would havehad to be found.

Laurence then looked at 14 different minicomputerinstallations and found he favoured the DEC PDP-11with RSTS operating system.

"Fortunately, we held our horses. Firstly, thissystem could not be built a step at a time. We wouldhave had a lot of spare capacity during the devlep-ment stages. Also at that time economic conditionsand unemployment in all sectors were not conduciveto an increase in capacity.

"Another factor of course was the amount of floorspace required to house the machine. A mini iscertainly smaller than a mainframe but it is by nomeans small. Then of course we weren't too surethat we wanted one system to handle all ourprocessing; if it went down our whole businesswould suffer.

"The next thing we looked at was kit computersbut rejected this idea as we couldn't get amanufacturer's maintenance contract on them.

"At an exhibition at the US Trade Centre I saw theDTC Microfile for the first time and met thecompany's president, Roy Worthington. It seemedthis machine could do what was wanted. Althoughthe DTC box only allows for two terminal ports,expansion facilities have been built in by EXTEL for afour port device to enable connection of more thanone Microfile together. This allows for redundancywhen one machine goes down; a facility not availableon the PDP-11.

"Using this machine we will be able to fullycomputerise three of our four agencies for a totaloutlay of about £20,000. With a minicomputer -basedsystem, it would have easily cost double theamount."

Computech is now in the process of expanding itsconfiguration with a second processor containingfour disk drives, and replacing its VDU with fourInformer VDUs. Informer did not provide 80x24upper and lower case screens previously butComputech has just received the first lot direct fromthe U.S. manufacturers. Laurence Payne expectsthat the combination of Microfile processors withpolled terminals will give him more computing powerthan with a mini system.

"As for other microcomputers, we did not reallyfind anything that provided dual disk drives andBASIC at that price at that time. As it turns out nowwe have had a company order a similar kit and ask usto do their programming for them. I think this kit canbe pretty competitive as a lower -end key -to -disksystem with Redifon and CMC systems."

With the dual processor configuration it will bepossible to have one spooled for printer output whilethe other will be free for processing independently ofthe printer.

"We've been so successful with this machine thatanother employment agency has now approached usto install a similar system on a turnkey basis. TheINFORMER terminals are ideal for this type ofapplication. The screen contains the full 24x80characters but is much smaller than average and sitson a stem to the side of the keyboard, which makes iteasy for an interviewer to enter particulars ofapplicants without having to crane her neck over thescreen."

This is all quite an impressive record consideringthat the hardware was bought only in November1976 and first live running was already under way inApril 1977: from do-it-yourself user to systems housein under eighteen months!

40


MK14-the only low-costkeyboard -addressablemicrocomputer!The new Science of CambridgeMK14 Microcomputer kitThe MK14 National Semiconductor Scamp basedMicrocomputer Kit gives you the power and performanceof a professional keyboard -addressable unit -forless than half the normal price. It has a specificationthat makes it perfect for the engineer who needs tokeep up to date with digital systems or for usein school science departments. It'sideal for hobbyists and amateur electronicsenthusiasts, too.

But the MK14 isn't just a training aid.It's been designedforpractical performance,so you can use it as a working componentof, even the heart of, larger electronicsystems and equipment.

MK14 Specification* Hexadecimal keyboard* 8 -digit, 7 -segment LED display* 512 x 8 Prom, containing monitor

program and interface instructions* 256 bytes of RAM* 4MHz crystal* 5V stabiliser* Single 6V power supply* Space available for extra 256 byte

RAM and 16 port I/O* Edge connector access to all data

lines and I/O portsFree Manual

Every MK14 Microcomputer kit includes afree Training Manual. It contains

To: Science of Cambridge Ltd,I 6 Kings Parade, Cambridge,

Cambs., CB21SN.

Please send me an MK14 StandardMicrocomputer Kit. I enclose cheque/Money order/PO for £43.55 (£39.95+ 8% VAT and 40p p&p).Lllow 21 days for delivery.

operational instructions andexamples for training applications, andnumerous programs including math routines(square root, etc) digital alarm clock,single-step music box, mastermind andmoon landing games, self -replication,general purpose sequencing, etc.

Designed for fast, easy assemblyEach 31 -piece kit includes everything youneed to make a full-scale workingmicroprocessor, from 14 chips, a 4 -partkeyboard, display interface components,to PCB, switch and fixings. Further softwarepackages, including serial interface to TTYand cassette, are available, and areregularly supplemented.

The MK14 can be assembled by anyonewith a fine -tip soldering iron and a fewhours' spare time, using the illustratedstep-by-step instructions provided.

Tomorrow's technology - today!"It is not unreasonable to assume thatwithin the next five years . . . there will behardly any companies engaged inelectronics that are not using micro-processors in one area or another. "

Phil Pittman, Wireless World, Nov.1977.

Name

Address (please print)

The low-cost computing power of themicroprocessor is already being used toreplace other forms of digital, analogue,electro-mechanical, even purelymechanical forms of control systems.

The Science of Cambridge MK14 StandardMicrocomputer Kit allows youtoleam moreabout this exciting and rapidly advancingarea of technology. It allows you to useyour own microcomputer in practicalapplications of your own design. And itallows you to do it at a fraction of the

_price you'd have to pay elsewhere.Getting your MK14 Kit is easy. Just fill in

the coupon below, and post it to us today,with a cheque or PO made payable toScience of Cambridge, And, of course, itcomes to you with a comprehensiveguarantee. If for any reason, you're notcompletely satisfied with your MK14,return it to us within 14 days fora fullcash refund.

Science of Cambridge Ltd,6 Kings Parade,Cambridge,Cambs., CB21SN.Telephone: Cambridge (0223) 311488

Science of I,Cambridge]

41


THE ENTREFRENEMR5

Ten minutes after we arrived at Computer Workshop,London, photographer Yoshi Imamura discovered he had afaulty camera and rushed off for a replacement. So the restof us went upstairs to a little office to talk, with thesecretary giving a busy and entirely appropriatebackground to the PCW interview with John Burnett andGordon Ashbee of Computer Workshop.PCW Creation stories always begin at the beginning . . .

Burnett We opened for business on 1st May 1976.PCW How did you get together?Burnett We'd known each other for eighteen monthsbefore. I called on Gordon as a representative and it wenton from there. Over drinks we would build electroniccastles in the air. I was then working for ComputerServices Centre in Acton, London, and computing hadbeen Gordon's hobby for three or four years.PCW Give us an idea of your backgrounds.Ashbee I did National Service as a fighter pilot in theRAF, worked for Coty, was a representative, then in 1958started my Entryphone business. I had to be independent.Burnett I did articles as a clerk in a firm of charteredaccountants, then trained for two and a half years in acomputer bureau. After that I had a short spell with ICL,

then joined the Computer Services Bureau as a

representative. I was initially involved at all levels - I likedthat. But later the usual job routine took over, I foundmyself a join in a rigid structure, a set-up which ofnecessity led to a division of labour. Eventually, not for me.PCW When did you connect with Southwest TechnicalProducts?Ashbee We were ordinary customers at first, but wehappened to be quite big ones. From there it was only astep to arranging a franchise. Incidentally, there is now afactory in Peterborough which is a joint venture with theAmericans but separate from Computer Workshop.Burnett The idea of starting Computer Workshop wasGordon's. I began by getting business from him and endedby getting into business with him.PCW What do you think of the hobby market?Burnett Ah yes, the hobby market. Is there a genuinefuture for it? I'd be able to answer the question better if I

could identify the hobby market. What is it? Where is it? Atthe moment it seems to me that it's subsumed in a

company ordering a small system and whoever using itindulging his hobbyist inclinations together with bread andbutter work. It's different in the United States because

42


there the level of personal disposable income is higher andthe tax structure different. Over here I'd say that themajority of so-called amateurs are trying to make abusiness out of their hobby.Ashbee I agree. In the US the amateur is truly anamateur. In Britain - up to now at least - people don'twant to spend money on a system without regarding it asan investment which must bring a return - in moneyterms.Burnett Here the genuine amateur will go for a genuinelylow cost system - a barebones. In America he is as likelyas not to wallow in expensive peripherals.Ashbee Think about it. Systems with 24k of memory,floppy disks with sophisticated operating systems: inreality, general purpose systems, sometimes with a lotmore power than the user has the time or ability orenterprise to use.Burnett And the amateur over here can't really substitutecassette for disk. Cassettes are a lot more awkward, a lotless flexible.Ashbee Mind you, I see more dramatic falls in prices ofperipherals - except for printers. There's a big marketawaiting the first person who comes out with a reallycheap and effective printer, one retailing at under ahundred pounds. Even floppy disk prices are falling. Forexample, drives are about two thirds the price they were ayear ago. The problem as always is with the mechanicalparts. Their prices tend to remain static.PCW How about the education market?Burnett It's very large and it's coming to us.Ashbee Computer education is percolating down into theschools.Burnett Many teachers are eager to introduce personalcomputing to schools. There was a case where a sciencedepartment bought a kit out of its own pocket.Ashbee The future of personal computing - in the homeand business - is being created now in the schools of ourcountry. What the future will be depends on how sensitiveeducation authorities are to the computer revolution, howflexible and farsighted they are.PCW Can we now get down to business?(The question was auspicious. Yoshi returned with aneager new camera began takinophotographs).Burnett Computer Workshop business with businesses isbrisk and increasing rapidly. To give you an idea, in thevery early days a businessman with no computingbackground rang us up, more out of curiosity thananything, to ask if we could help in processing a largenumber of forms which needed forty minutes each inhandling manually. It seemed to me like a thirty lineprogram in Basic. I took the details, wrote the programovernight, then he came over for a demonstration. Thefollowing week he ordered a system. All this time he'sbeen using just that one program and is still saving money,effort and time. Now that he's more aware of the full

power of the SWTPC 6800 he wants to go on to wordprocessing. We've suggested adding a floppy disk andhe's in line to upgrade his system so that it's generalpurpose.Ashbee I might point out that in education people aremore concerned with systems software, because manyschools and nearly all universities are perfectly capable ofwriting their own applications software, a lot of it brilliant.But in small businesses there's a need for applicationssoftware. This offers profitable opportunities for smallsystems houses and even free-lance programmers.PCW Lets go on to customer relations. Are you on theside of the angels?Ashbee We try to be! We have very few complaints oneither the hardware or software side. We do haveoccasional complaints about delivery.Burnett Now that the factory in Peterborough is a fact,complaints should get even fewer. It's hard for people torealise the difficulty we've had setting delivery dates. Ineach system, there are roughly two hundred and fiftydifferent types of components from various manufac-turers. It only takes a shortage of one of them to affectdelivery. Which is not to say that we ourselves don'tsometimes make mistakes.

In the early days we had friendly relations with each ofour customers - we were small enough then. It paid offbecause our reputation is founded on the soundness of theSWTPC 6800 and this goodwill. Now we're expanding at aterrific pace - currently at nine per cent a month,cumulative. We do have a self imposed restriction ongrowth since we're working at full capacity and not adver-tising indiscriminately.Ashbee It's a tribute to the product.Burnett I really regret that now we don't know everycustomer by name. With the best will in the world, wecan't. When customer relations become more formal theyare in my opinion not so good. But we do still give indi-vidual attention to someone who has a problem. We doour best and we're proud of this.PCW So, do you regard yourselves as amateurs inbusiness?Ashbee Perish the thought. Our approach is professional,though we have an amateur's love of computing. Let's saythat one does one's best under intense pressure.PCW Finally, what do you think of possible Japaneseinroads into the personal computer field?Burnett Not what but when. It's only a matter of time.The Japanese are currently investing hundreds of millionsof dollars in computing in all its aspects. It is Japan'sintention to become the world's first information society,ahead even of the US. As other markets run down or arerestricted the Japanese are turning to one of the biggestgrowth areas of all time - computing. Their impact is sureto be felt in personal computing even if only as a by-product of a major effort in big computers.

43


PET 2001 Reviewby John Coll

PCW The PET is a very attractive small computer in appearance and design. It is a successfulattempt at placing computing within the reach of thousands and there seems little doubt thatthousands worldwide will buy it. Thus features which now seem to be non-standard will come to beaccepted and it is probable that software and hardware which is CBM-compatible will be offered inthe same way that there are now alternative choices to IBM software and peripherals. PCW

Let me start by giving you my overall impression ofthe PET 2001:- it is a very well constructed machinewith many nice features but I was unable to spendmore than a day with the machine: a week wouldhave been more appropriate.

The machine is well built, solidly constructed andhas a bright sharp V.D.U. display. The internalconstruction is very clean - a single printed circuitboard carries all circuitry except the power supplyand monitor scan circuits. Connections to theoutside world are made via 4 edge connectorsdirectly on the main p.c. board. I found the calculatorstyle keys on the keyboard too small to usecomfortably at speed. The integral cassette driveresembles an ordinary cassette transport but it wasquite impossible, in the time available, to examine itin detail.

The keys normally produce upper case outputon the screen, but if the "SHIFT" key is depressedthen the displayed character is a graphic. Thegraphics characters are not of the usualViewdata /Ceefax/Oracle "3 squares by 2" formatbut instead are generated by a "graphics ROM". Thiscan generate such characters as "hearts" and"diamonds" as well as a good range of lines andshaped area from which complex pictures can beeasily built up. This is a very interesting departureand makes picture -drawing very straightforward.The lack of compatibility with other systems maywell be a price worth paying.

Getting into lower case, however, is not so easy.The command is "POKE 594.68,14". Havingexecuted this command the display will show upperand lower case letters but not graphics characters."POKE 59468,12" enables upper case and graphicsto be displayed but not lower case letters. That is apretty inconvenient facility. POKEing 59468 withanything else usually causes the system to crash withtotal loss of program and data - but fair enough,one expects the command POKE to wield havoc ifused carelessly.

The BASIC interpreter is resident in Read OnlyMemory - which is convenient in some ways but itdoes mean that if (when) any bugs are found orextensions to the language are needed there will be

nothing that you can do except buy a new set ofROMs from CBM, and that could be very expensive.It also means that if you find a better version ofBASIC available from someone else you will not beable to put it in without buying a whole lot morememory from CBM. The interpreter has been writtenby Microsoft who supplied Altair with their BASIC.The timings of those benchmark programs whichwere given in the last issue are:BM1 1.5 sBM2 9.7BM3 18.2BM4 20.3BM5 21.8BM6 32.4BM7 50.9BM8 12.2BM7+8 63.1

The really outstanding feature is the screen editfacility. (See Patrick Sutton's article in this issue -Ed.) One can move the cursor to any line of programthat is displayed on the screen and with singlekeystrokes one can insert, delete or change anycharacter. On hitting the RETURN key the new line isinput to the interpreter. Thus small errors are veryeasily corrected. One can also use this feature toenter a series of identical lines with different linenumbers. To do this one enters the line in the normalway and then edits the line number before hittingRETURN each time. That is a really useful facility.

There are two major omissions in the software.The first is the lack of a RENUMBER command. Yes,one can manage without it, one can manage withoutalmost any particular command, but anyone who hasused an interpreter with RENUMBER will know thatit is vital. (Do you remember that strange idea thatone heard once, that Computers were meant to saveyou time? No . . perhaps I imagined it.) The othermajor omission is the impossibility of stopping aprogram in mid -stream, changing a line or two andthen continuing with variables intact. If ANYchanges are made to your program then all variablesare zeroed. When a program is saved the status ofvariables is not automatically saved - so if youintend doing a two hour run then you had bettermake sure that your own "variable save" programworks first - and don't decide half way through thatyou want to save one of the other variables!Strings are limited to 256 characters and there is noPLOT command to match that found in other inter-preters. It is not possible to tell BASIC to leave asection of memory free for machine languagesubroutines. BASIC stores strings downwards fromthe top of memory and BASIC program text andvariables, upwards, so you have to take a leap intothe dark when trying to place any machine languageroutines in memory. There is one 192 byte block ofmemory available between $033A and $03FF. TheRandom number generator is not a True RandomNumber Generator, as stated in their white leaflet,but is the usual pseudo -random sequence generator.It seems to work well. One slightly awkward featureis that although the machine works in ASCII

44


ii rw, o#

(DILu o%

y&

r\

"---(

( inLL,

Fillj - jIirQ ,

fo,itife,1W E t.li

oliff-2-1E'TJ l' 0(.1

770

ollinPJL4

4,d -71(g,T

E 'a 7 ) T( (I 7 'AURN

.--4-Z

'--4-'XS_JVE 1 X 8NJJ-T--1M

O N..-:--.

(SHIFT 7'1fIVS ( H '0

33C

M] SPACE

<

ril In i 'RUNSTOP

r-C/R

MN. .

(''MSSky

CC'S""

(IMODEL )1

7 8 9 / J

II' MI l 7-1' 7-114 1 5 J 6. .)I(- J

"Tc'l1

FIT'2

IT3

CB

+

1.-Trs --ET fi- =

The keyboard

internally the V.D.U. does not store alphanumericswith their normal ASCII code. A is stored as 01, B as02 etc. down to Z stored as 26. Zero to nine arestored with codes between 48 and 57. That seemsreally strange since the ASCII code is one thing thateveryone understands, and it surely would have beenjust as easy to put the right code in the right place inthe ROM. This is going to make interchange ofprogram material difficult.

The machine is certainly individual: non -CUTSstandard cassette, BASIC in ROM, non -ASCII VDUdisplay, IEEE interface not plug compatible; and they"PEEK protect" the screen editor ROM so you can'tuse the disassembler to look at the routines they useto convert the internal ASCII to their own VDU code.

Talking of disassemblers brings us to themachine language environment. Well, quite simply,there wasn't one. There was no machine languagemonitor in any form - nor was data of any systemmade available. I understand that the KIM monitor ison its way. There is no editor (in the accepted senseof that word) or assembler. I did use their dis-assembler which was written in BASIC. I was toldthat it would be available for £10 and that seemsreasonable if it proves to be so. It is quite adequatethough it did not attempt to label statements thatwere referenced elsewhere or attempt to deal withtext strings in any way. I did not have enough time totest it thoroughly, though. As I have said, areas ofROM are PEEK protected and therefore cannot beimmediately disassembled. The protected area runsfrom 49152 to 57599 ($C000 to $E100) and includesthe whole of BASIC, and at least some of the screeneditor routines (if their memory map is accurate).Clearly anyone working with machine language textroutines will need to use similar (identical?) routines.He will therefore need to copy the ROM programs toan area of memory that is not PEEK protected andset the disassembler going on it there - as anattempt at protection it is totally ineffective, andcould be an irritant.

At present there is no floppy disk system or hardcopy unit and C. B.M. are not at present giving anyindication of when they are likely to be available.Anyone who has used a computer will know thatthese two are absolutely essential in the long run. I

will be very interested to see the pricing of theseitems. Will they be as much of a bargain as the PETitself is, or will they carry a substantial price tagwhich the already committed PET owner will have topay? Anyone want to place bets? I think prospectiveusers would be well advised to wait and see.

The cassette system uses a non CUTS standardformat, which is unfortunate, and in the time

available I was unable to see if it would be easy towrite machine language software to read and createCUTS format tapes. It is not clear from the intro-ductory booklet whether the second cassette thatthe system can support needs to be special in anyway. One would hope that an ordinary audio cassettemachine could be used unmodified. One quote fromtheir glossy leaflet: "A second cassette is availablethrough a built-in interface. This allows true filekeeping and places the PET as a true andconventional electronic data processing systemequivalent to just about any large commercialcomputer". That statement is, in my opinion, utterlymisleading. The access times (for file maintenance)are vastly different.

The cassette file handling system has theunfortunate facility that file name matching can workon substrings. Thus if the command LOAD "CAT"were given the file handling routine would load in anyof the following files "CATALOG", "CAT3" or"CATCH". The lack of a tape counter or of anymeans of fast -searching a tape - features found onother systems - are serious omissions. I was unableto establish whether binary files could be saved, reador appended and whether or not a "link" addresscould be associated with such a file. The cassettesystem is pretty much a bare minimum. It was alsonoticeably easy to "crash" the PET into a state fromwhich recovery was only possible by removing themains plug with resultant total loss of program anddata. A "RESTART" switch is essential. Even if asystem "crashes" it is often possible to recovermuch, or all, of your program if an orderedRESTART is provided.

It is fairly easy to form an impression of themerits and demerits of machines and their software,given adequate opportunity. What is far moredifficult, but very very important for the prospectivecustomer, is to weigh up companies' attitudes totheir customers. The system that you own willexpand. You will want new facilities as they becomeavailable, you will want to expand memory as itsprice decreases, you will want to run competitors'software and to write your own routines. You willhave problems and need help from companies. Tosuggest otherwise is naive. When purchasing yourfirst substantial chunk of computer you will want toconsider companies' previous records. Have theybeen prepared to help with your difficulties? Havethey offered software freely at reasonable prices withgood documentation and listings of vital parts suchas the input-output routines?

These are questions worth asking and worthgetting answers to.

45

313 Kingston Road, Ilford,Essex, IG1 1PJ England.01-553-1001

zko

NTERPRISESFrom the representatives in Europe . . . for America's leading Micro -computer magazines and books, for thehobbyist, educationalist and professional alike, we bring you a little light BrowsingReading maketh a full man . . . Francis Bacon (1561-1626)

TICK HERE OR INDICATE QUANTITY ORDEREDFrom Adam Osborne Associates

INTRODUCTION TO MICROCOMPUTERSVolume 0: The Beginners BookVolume 1: Basic ConceptsVolume 2: Some Real Products (Revised Late 1977)

6800 Programming for Logic Design8080 Programming for Logic DesignZ80 Programming for Logic Design

(Available May 78 approx)8080A/8085 Assembly Language ProgrammingSome Common BASIC Programs

BUSINESS PROGRAMS IN BASICPayroll With Cost AccountingAccounts Payable & Accounts Receivable

(Available from June 78)General Ledger (Available June 78)

From Scelbi Computer Consulting Inc.6800 Software Gourmet Guide & Cookbook8080 Software Gourmet Guide & Cookbook (updated 78)8080 Programmers Pocket Guide8080 Hex Code Card8080 Octal Code Card8080 Guide and One 8080 Code Card8080 Guide and Both Code CardsSCELBAL High Level Language for '8008/8080' SystemsSCELBAL String Handling SupplementSCELBAL Extended Maths SupplementUnderstanding Microcomputers & Small Computer SystemsSCELBI 'BYTE' Primer8080 Standard Assembler (In Book Format)8080 Standard Editor (In Book Format)From Peoples Computer CompanyReference Book of Personal & Home ComputingWhat to Do After You Hit ReturnDr. Dobbs Journal Volume 1*From Kilobaud/73 Magazine Inc.Hobby Computers Are HereNew Hobby Computers

From Dymax Inc.Instant BASIC by Jerald R. BrownYour Home Computer by James WhiteMy Computer Likes Me . . . When I Speak

BASIC By Bob AlbrechtGames With A Pocket Calculator by

Thiagarajan & StilovitchGames, Tricks and Puzzles For a Hand

Calculator by W. Judd

*From BYTE Publications Inc.Paperbytes:

Tiny Assembler for 6800 SystemsBar Code Loader for 6800, 8080, Z80 & 6502 Micros

Best of BYTE Volume 1

£5.95£5.95

£11.95

£5.95£5.95

£5.95£6.95£5.95

£9.95

£9.95£9.95

£7.95£7.95£2.25£2.25£2.25£4.20£6.00

£39.25£8.00£4.00£7.95£9.95

£15.95£9.95

£4.95£7.00

£10.00

£3.95£3.95

£4.95£4.95

£1.65

£1.75

£2.49

£5.75£1.75£8.95

HOW TO ORDERPlease note our prices includepostage and packing, but notinsurance, if wanted add 12pfor every £10 of booksordered. Make cheques, PO'setc payable to:-

L.P. Enterprises, Dept. PBCREDIT CARDS accepted:BARCLAYCARD/VISA/ACCESS/DINERS CLUB/AMERICAN EXPRESSPhone 01-553 1001 for all

credit card orders (24 hourservice)

TICK HERE OR INDICATE QUANTITY ORDERED

*From Creative Computing Press

PriceUK

PriceOverseas

If Different

Best of Creative Computing Volume 1 £ 6.95Best of Creative Computing Volume 2 £ 6.95101 BASIC Games (Revised & Reprinted

Feb. 78) £ 5.50The Colossal Computer Cartoon Book £ 3.95Computer -Rage (A new Board Game) £ 6.95Artist and Computer £ 3.95Three Binary Dice £ 1.00*From Everyone ElseTV Typewriter Cookbook by Don Lancaster £ 7.95Magazine storage boxes (Holds 12 minimum) £ 1.75Sybex: Microprocessors £ 7.95Sybex: Microprocessor Interfacing

Techniques £ 7.95Dilithium: Home Computers

Volume 1: Hardware £ 6.50Dilithium: Home Computers

Volume 2: Software £ 5.95

MAGAZINES: Back IssuesPersonal Computing £ 1.75Interface Age £ 2.00Dr. Dobbs Journal £ 1.60Computer Music Journal £ 2.50Peoples Computers £ 1.50BYTE £ 2.25Creative Computing £ 1.75Calculators & Computers £ 1.75ROM £ 1.75Kilobaud £ 2.2573 £ 2.00

MAGAZINES: SubscriptionsPersonal Computing (Twelve Issues Yearly) £16.00 £17.00Interface Age (Twelve Issues Yearly) £20.00 £20.50Dr. Dobbs Journal (Ten Issues Yearly) £13.00 £13.50Computer Music Journal

(Four Issues Yearly) £ 8.50 £ 9.00Peoples Computers (Six Issues Yearly) £ 8.00 £ 8.50Kilobaud (Twelve Issues Yearly) £20.00 £21.00*BYTE (Twelve Issues Yearly) £15.00 £15.00Creative Computing (Six Issues Yearly) £ 8.50 £ 9.00Calculators & Computers

(Seven Issues Yearly) £10.00 £10.50ROM (Twelve Issues Yearly) £16.00 £17.0073 (Twelve Issues Yearly) £20.00 £21.00

PCW2

Send to address above for the attn. of Peter, Dept PB All Orders must be PrepaidIndicate Payment Method:

My cheque, P.O., I.M.O. is enclosed in Sterling on U.K. BankCharge to Barclaycard/Visa/Diners/American Express/Access

Credit Card number Expiry Date

Name

Address

POSTCODE

Total Enclosed £

Card holders signature

All publications are published in U.S.A. and shipped air -freight by L.P. Enterprises. In unusual cases, processingmay exceed 30 days.*BYTE Subscriptions are processed in the U.S.A., and are air -freighted then posted from Amsterdam, allow 3 months for sub to start.

TRADE ENQUIRIES WELCOMEYou may Photocopy this page if you wish to leave your PCW intact


PCWREVIEW

Mike Dennis

NIESE115011111113191111E5 315

About the beginning of 1977, various advertisements appeared under the Sintel banner for theRML 380Z microprocessor systems. In fact, it was not until August 1977 that the first prototypemachines went out and December 1977 that the first production models were despatched. The pricesranged from about £400 to £1400 depending on facilities and memory size and it was the grand -daddyof the range with 32K of RAM that I was given to review.

Originally I was to have it for only a week but I must have smiled nicely for Research Machines(RML) were quite happy for me to have it for a much longer period. In fact, I wholeheartedly endors,ethe editor's comments about the interest and assistance offered by this firm. Such attributes are alltoo often non-existent in larger organisations.

Appearance can be deceptive!Surprisingly, the 380Z is much lighter than its' sizesuggests. It weighs only about 251bs. in a cabinet thesize of a small suitcase. No wonder, really, as whenyou open up the case all that you see are two PCB's,a power supply and a lot of air! (See photograph 1).

The 380Z was designed to be modular in construc-tion, which is good. What this means is that faultfinding is greatly assisted and spare boards or powersupplies can be swiftly changed and thus reduceyour downtime. What I didn't like was the use ofLucar connectors on some wires which were far toothin to carry the relatively heavy weight of the con-nectors. There are many other more suitable con-nectors on the market and so to use this particulartype is sloppy. The general standard of constructioninside was, I felt, more adequate than elegant. Forinstance, odd pieces of perspex sheet were stucktogether in layers and then stuck to the inside ofsome of the panels presumably to stop them fromrattling. The actual fixing flanges in the internalpanels distort permanently with use as the screws aretightened up. You can see this in the photograph.Generally, in the professional equipment one doesn'tnormally find this. Surprising, too, was the lack of

Serial Number or other means of individualidentification.

The power supply has been mounted well out ofthe way to give plenty of room in the case. Morespace has apparently been saved in this area by usinga transformer designed by RML. All mains con-nectors were suitably protected against 'fairyfingers'. The mains input itself is via the detachablenew IEC connector and comes ready to go with amains lead of decent length and 13A plug. I musttake RML to task, though, for not fitting a fuse oflower rating in the plug. I would have thought thatone at 3A was more than ample and not 13A! (Dearreader, are all your plugs correctly and, therefore,safely fused?)

Beautiful PCB's

Another grey mark is the lack of protection for thetwo PCB's. They rest on the bottom of the casedirectly and although they are free to slide up anddown, they are still vulnerable to shock or impactthrough the bottom panel. Perhaps if RML agreethey could fix some form of foam plastic to thebottom of the case to rest the boards on? The PCB's

47


themselves are beautiful - there is no other word forit. Of German manufacture, they are double sidedand have about 1400 plated through holes each. Thepair that I had were slightly spoilt by solder splasheswhich can cause intermittent faults. RML say thatthey hadn't been cleaned properly and certainly otherboards that I saw were much better. In any case, I

am really nit-picking as since the boards are mountedvertically then any solder splashes that fall offshouldn't have any catastrophic effect.

The physical size of the boards is 10" x 51" and that'sthe nearest that RML get to the S100 bus - the samewidth! Provision has also been made for mounting ashorter six inch board. I think their bus techniqueparticularly cost effective. They make use ofconnectors similar in style to Scotchflex but made byITT Cannon. Essentially, they consist of a male(female?) half mounted onto the PCB and the otherhalf crimped onto a flat ribbon cable. Extension leadswith any number of plugs/sockets can be made up inorder to connect up the various busses to any extraboards. Mind you, as the CPU board canaccomodate up to 32K it would be some time,probably, before you needed any extra boards formemory! It was a pity that no indexing was providedas in theory it is possible to insert the plugs eitherway round. In practice, the physical layout andwiring prevents you from doing this. You can do it,though, with the dc supplies which is potentiallydisastrous. You could end up with a very expensivemistake on your hands. Take note!

Connections to the outside world are by a sevenpin DIN socket and two sockets for a visual display.RML supply one DIN lead and one co -ax lead forconnecting up. Provision has already been madeboth in the wiring, hardware and software to coupleup two cassette machines via a suitable splitting boxto the DIN socket (see later). I was not supplied withany of this but have seen a working system. Clearly,an indication that RML are continuing to expand thesystem, which is to their credit.

SignalsAlthough both sockets for the VDU were Belling

Lee type co -ax, only one socket is fed by the UHFmodulator. In practice, there was sufficient stray

2

oscillation coming from the second socket to provideanother UHF signal for a second TV. However, underthese conditions it is possible that some TV's couldinteract between themselves to produce inferiorpictures. RML never intended it to be operated in thisway but nevertheless you could get a small bonus.

I tried feeding the UHF signal into a colour set andhad great difficulty in tuning in the signal. Now most,if not all, colour sets have an AFC circuit which youdisable when you tune in a station. When a similartechnique was tried with the 380Z, as soon as theAFC was enabled again, it pulled the set off tune andso produced a bad picture. But if the AFC waspermanently disabled by going inside the set, theproblem then ceased as one would expect. Oninvestigating further, the modulator was giving out asignal of about 10mV! More than enough to swampany normal AFC circuit and so cause the de -tuningthat I had experienced. The simple cure is either touse an aerial attenuator (available from any good TVdealer) or use a black and white set. However, thishigh signal level is rather naughty as the frequency ofthe modulator was about 592MHz or Channel 36 andso outside the broadcast band. There are other usersaround this frequency and so at signal levels of thisamplitude it is conceivable that interference could becaused by the 380Z to any authorised users of thisband.

The second socket is for a line -fed monitor and assuch should have been fitted with the proper socketfor the job i.e. a BNC (readily available fromRadioSpares). Belling Lee co -ax sockets aredefinitely not professional for this type of work and inany case necessitate a convertor lead for use withany monitor. Murphy's Law applies when looking forthis lead. Again, the video output level is a non-standard one of 0.15v pk-pk of syncs and 0.4v pk-pkof video. (If your system measures differently thenthis is due to production tolerances). I measuredthese when the signal was terminated (as it shouldbe) in 75Q. Leaving the signal unterminatedincreased the amplitude but made the characterslook a bit ragged, which was expected. Thedefinition in the monitor was, of course, much betterthan that obtained from the UHF feed. This isgenerally the case and I wish that better modulatorswere readily available.

48


3

HardwareTalking of volts brings us conveniently to a brief

look at the actual hardware used, as I expect thatsome of you may be particularly interested in thisaspect. As can be seen from the photographs 2 and3, there are remarkably few IC's for so versatile amachine. There are approximately 45 ic's per boardtogether with their respective sockets. I could writean entire article on the various and neat designaspects of the 380Z but suffice to mention just a few

1) 16k x 1 350nS dynamic RAM's were provided.They are NEC/Nippon devices and similar to theMostek 4116 etc. As such, 4k x 1 dynamics are adirect replacement fit and provision has been madefor this.

2) Thick film resistors have been used where anypull-up and pull -down resistors are required.

3) An 8MHz crystal is used for the system clockand divided down to provide a 4MHz clock of equalmark -space ratio.

4) One Wait state is put in all Memory Request(MREQ) cycles at the factory. Should you wish torun at a slower rate then the user can re -arrange on-board wire links to insert a Wait state in each M1cycle.

5) Much use has been made of LSI devices such asthe 74LS393 (two 7493's in one package). There is aliberal sprinkling of 74LS157, 158, 241, 244 and 374's.The majority of these are comparatively new devicesto the TTL family and many have been designed withthe microprocessor in mind. If you are at allinterested in designing your own system then I canonly urge you to buy an up-to-date TTL Data Bookfrom one of the distributors. They make fascinatingbed -time reading!

6) Liberal decoupling means gbod noiseperformance. RML have spread tantalum capacitorsand 0.1uF's around like confetti and the result hasbeen very little noise on any of the dc lines. Mostcommendable. Even the 3A regulator which issituated away from the CPU board by a very longwire seemed perfectly able to cope. In fact the 380Zonly uses about 1} of the available 3Amps whichagain allows ample expansion.

7) The COS (Cassette Operating System) lives intwo 2708 ROM's and there is space for an optional

ROM be it user programmed or as, in my case,supplied by RML (more anon).

8) I spied a couple of other ROM's on both boards.On the CPU board were both a 74S287 (256 x 4 bits)and a 74S288 (32 x 8 bit) ROM. The 287 is used todecode which block of RAM is being addressed andthere are four versions available depending on whattype of memory you have leg. 2 x 16K, 2 x 4K etc.).Should you win the pools or strike lucky at thefinance meeting and wish to install more RAM, thenall you do is swap the 287 and plug in another RAM.Slight alterations to one pin have to be madedepending on whether you are using 4K or 16KRAM's and these are done on wire -wrapconnections. Should you have 8K of RAM and wishto update to a 16K version then rather than supply aseparate board as originally advertised, RML will buyback your 8K of RAM and sell you a 16K block. Anyextra cost doing it this way will be absorbed by them.

One output of this ROM connects up to the 'S288as an enabling signal and indicates that the lower 8Kbytes of memory are being addressed ie up to 1FFF.This and address lines A9-Al2 inclusive are thensufficient to decode the three ROM positions, theROM data buffer, VDU memory select signal and'I/O' port select signal. This latter when NANDedwith A2 -A8 inc. provides a low when locationsOFFC-OFFF are being addressed. However, I/O isreally a misnomer as the IORQ signal from the CPU isnever used as these particular I/O ports are memorymapped. OFFC is either the keyboard in the readmode or Port 0 in the write mode. The pattern of bitswritten into this port control various functions in themonitor. OFFD is a counter used for the softwaresingle step and reset. OFFE is Port 1 and dedicated tothe cassette/program storage facility etc. and othermisc. functions. OFFF is a user port and has evenbeen provided with a buffered octal latch.

So with only two ic's, RML have decoded all ofthis whilst still maintaining a high degree of flexibility.A really good example, I think, of what a freshapproach can do in a situation like this. That reallysums up the CPU board.

The VDU board contains both VDU and enoughinterface for two cassette recorders. Again,ingenious use had been made of two ROM's ('S287sagain). These, in conjunction with their respective

49


line and field counters, provide all the sync andblanking waveforms, reset signals etc. required bythe VDU together with all the addresses required bythe display RAM's and charactrer ROM. This is thefirst time I have seen such a clever design. The syncwidth is a little non-standard (4uS as opposed to4.7uS) but in practice this should make very littledifference.

Any calls to output an alphanumeric character tothe screen occurs in Field Blanking. In order to usegraphics in an interactive mode a faster rate isrequired. One sub -routine in COS uses line blanking(24uS) to output one character per line. Now, 24uSis not long, even with a 4MHz clock so if you wait forline blanking to arrive before beginning your o/proutine then you soon run out of time. However,since this output routine consists of a known lengthof machine cycles and hence time, you cananticipate the lineblanking signal and use a pulse ofearlier timing. Because RML have done their syncgeneration the way they have, the line blanking ROMquite conveniently has just such a waveform with theright timing!

One design point that I personally dislike is thesoftware conversion of capital 0 to . This isnecessary because the Texas 262 charactergenerator used doesn't discriminate between capitalO and zero. I personally prefer 0 and 0.

Note from the photograph (2) the conspicuouslack of any UART's (a UART is a large 40 pin is thatcan convert 8 bit parallel data into serial data and viceversa). This is because about 256 bytes of COS hasbeen used to record and replay a cassette tape toCUTS standard. At 2708 prices this costs veryapproximately £2.50 whereas two UARTs could costtogether about £10 - clearly a good example of aneconomical software/hardware tradeoff. However,this is only the half of it as with the option ROMsupplied there is a fast cassette option of 1200 Baud.Bear in mind that their cassette file handling is alsovery sophisticated.

RecordingTo record a programme on tape you either type D forDump or SAVE depending on whether you are in theCOS system or have BASIC loaded in. Eitherprompts you with a request for a file or programmename. In COS, this can be up to six letters long; inBasic, ten. The system then prompts you with arequest for the first and last address of the data orprogramme and also the start address. Theprogramme is then parcelled up into either 128 or 255(yes, 255 as in order to save bytes in the monitor,RML use the zero count for end of data block) byteblocks. Each block has the programme nameattached to it.

LoadingTo load a programme from cassette, select yourrecording speed (300 or 1200) if necessary, enter theprogram name and then replay the cassette. TheCOS system reads in all the data blocks and echoesthe name at the beginning of each block on to theVDU screen. When the name coming off the tapematches the programme selected by you, themonitor loads the programme at the correct address.Since each block is named, either you or the 380Zcan search through an entire tape for a specificprogramme. All that you need to do after enteringthe programme name is to fastwind through thetape, periodically stopping and replaying a shortsection of tape to see in which programme you arein. Likewise, by using cassette recorders (such as theHitachi loaned by RML) with a remote pause meansthat the 380Z can load a block of data into itsAssembler (ready soon!), then pause the cassettewhile it digests the data. Of course, this latter facilityis only provided by their Cassette File HandlingProgramme which is supplied free with the moreexpensive versions. This programme can also beused, so RML say, for loading tapes recorded toother standards such as the INTEL format. Aninteresting side effect of the basic COS system is that

+PC 4400 32 D3 48 87 28 58 C5 CDSP 40FE FE 40 00 44 86 03 C3 81TY FFFF FF FF FF FF FF C3 38 03Ix 4100 00 44 86 03 C3 81 41 C3HL 0386 D3 00 F7 16 FE 4C 28 3FDE 0005 38 03 C3 70 03 54 4F C3BC 0000 FF FF FF FF C3 38 03 C3AF 1A4A Z H

43F443F543F643F743F843F943FA43F8

06CD9149C3774378

iti>4400'I

43FC 3243FD D343FE 4843FF 87

+4400 28+4401 584402 C54403 CD

4404 644405 464406 7D4407 FE4408 504409 30440A 4C4408 07

440C 38440D 49440E AF440F C84410 3D4411 304412 014413 3C

50


programmes recorded in this fashion have a certain (Iadmit, not fantastically great) security against easypirating.

I used the cassette system quite extensively and,at first, had one or two failures. This prompted a visitfrom RML with various spare bits but we were unableeither to reproduce the fault or to explain why theyhad occurred in the first place. Certainly, somefailures can be put down to my lack of familiarity withthe cassette machine. This is one of the few cassettemachines that has both volume and tone controls onthe output. Most cassette machines have neitherbut, in this case, it was necessary to adjust themcorrectly. One annoying feature of this cassetteplayer was that connecting the DIN lead mutes theinternal speaker. Personally, I like to have the volumeup a bit so that I can hear what's going on and so ifthis too is your preference then this is not thecassette player for you.

A more annoying feature was the automatic returnafter the sixth letter in the name when loading inCOS. If you make a typing mistake in the sixth letter,then you have to either Reset the 380Z or ignore themistake, replay some tape and then hit Control C.(This is because you cannot return to COS if the tapeisn't running. This is one drawback of using a

software cassette interface in this way.) With eitheryou have to go through the whole loading procedureagain.

Mind you, hitting the Reset isn't as disastrous as itcould have been had RML not had the forethought todo it carefully and keep the refresh going for thedynamic RAM's during reset! However, you've gotto remember to reset your fast tape option, ifnecessary, after reset, otherwise you'll enter yourname, start the machine and then have to start allover again! This is because COS 'fails' to normal tapespeed after a reset. A better way would be a 'fail' to'fast mode in' as most programmes will be stored ontape in this way.

One nice feature of the COS loading is that shouldyou get a read error then the machine won't abortcompletely, unlike some. You have the option ofeither aborting deliberately or of rewinding thecassette a bit, typing 'L' and the programme will bereloaded at the start of the next block, and hopefullygo in OK the second time round. I noticed that youcouldn't do this when Saving a Basic tape.

I tried playing back pre-recorded cassettes madeon various cassette machines and had nocompatibility problems. I measured the "dynamicrange" over which the COS would load and savetapes satisfactorily and found the system was stillcoping at a level of 20dB below peak recording level.What this means is that you should have no need toworry unduly about what level to record at whensaving programmes. Obviously, when using the highspeed option, tolerances regarding drop -out etc. onthe tape are that much tighter and RML quitesensibly recommend the use of high quality tapes.

One possible problem with some cassettemachines is that the output is 180° out of phase withthe input. RML have thought of this and provide apre-recorded programme called "TSTCS" or TestCassette. This enables a check to be made both foroptimum replay level (assuming you have control)and for phase. This is all fairly well documented inthe literature and any out of phase machines caneasily be cured by using a 1:1 transformer suitablywired.SoftwareAny system, no matter how good the hardware,must stand by its firmware and software and it is herethat the 380Z scores. In particular, the 'softwarefront panel' is one of the clearest and most versatilethat I have met (see photograph 4.). Briefly, itdisplays in the upper unscrolled 20 lines of thescreen, the contents of all the main registers, stackpointer and programme counter. Those registers thatcan be paired up are and the resultant 16 bit address

51


thus derived is displayed alongside. The contents ofthe memory location thus addressed by each pair isalso displayed together with the contents of severallocations on either side, the actual location beingpointed to by the arrows. Personally, I would havepreferred the display of these locations to stand outfrom the rest of the display (bright -up?) as it wasoften difficult to remember where each one was. Thecontents of the registers and programme counter canbe altered at will.

The data display, of course, will alter as theseregisters are changed and so the effect on the systemof any changes can be instantly seen. The stackpointer cannot which could possibly prove annoyingat a later date. Neither the I nor the R registers areshown. It is possible to switch to the 'other'registers. Another good point is the ability to singlestep through a programme. Pressing 'Z' causes anNMI into the single step routine. The programmecounter, stack pointer and registers all change beforeyour very eyes and one can really appreciate theeffects of the various instruction code, stack pointermanoeuvres etc. / can't recommend this featurestrongly enough for anyone involved in teachingmicro's. Go and see for yourself. A nice touch is theactual printing of the various flags when set e.g. Zand C if both the zero and carry flags are set.

All this display is contained in ten lines. Theremaining ten lines are taken up by a separatememory pointer. The address of this can be enteredby the user following an 'M' command. The pointerwill alter and 32 bytes of data on either side of thisnew pointer will be displayed. The pointer can alsobe incremented or decremented by either 1 or 8bytes. Data can be written into memory directly andthis is how a hex programme would be keyed in. Thebottom four lines of the screen echo keystrokes andscroll.

Another nice feature is the 'Memory Shift'command. This works out the best way to move thememory blocks so as not to corrupt the data shouldyou wish to move one block to an area that, in fact,overlaps part of the original block. Many othermonitors eg. Zapple will corrupt the data undercertain conditions.

One or two minor bugs were: -1) any non -hex character acts as a terminator

and in one case caused the system to lock -upnecessitating a 'Reset'. This means that the monitoris not very forgiving and that if you make a typingmistake, COS will not let you correct it! One man'smeat . . . though, as RML say that some peopleprefer a system this way because you don't have toalways hit CR to terminate. You pays yourmoney . . .

2) Confusion due to choice of keys for certaincommands eg 'Z' in 'frontal panel' single steps (a Gotype signal) but 'Control Z' in Basic halts theprogramme!

3) Good though the single step facility is, itwould soon get pretty tedious stepping through aloop with a count of 255. RML have thought aboutthis and included the provision for setting breakpoints. What can happen is that the breakpoint isnever reached. The keyboard is then locked out andso the only resort is 'RESET'. Unfortunately, as thiswipes the registers, tracing the bug can provedifficult. What is needed is some form of "time-out"facility to return to the monitor. Not many monitorsdo this and unfortunately COS is no exception.

COS resides at 0000-07FF (see memory map).Beware, there be dragons, in the form of the restartvectors. RML have used these to great effect. Fourare jump outs into RAM for user control. One is the

normal reset to 0000, one is used for break points andanother for simulating RELATIVE CALLS. (see later).By far the most ingenious, however, is F7 which is aTrap Call and is a two byte instruction equivalent to a'CALL'. The second byte following F7 contains thetrap code of the desired target sub -routine. If all thecommonly used sub -routines eg. Get keyboardcharacter etc. are routed through as 'transfervectors' using this Trap Call, then the number ofbytes saved by not having to use a 3 byte Call everytime will more than offset the bytes needed for theTrap Call routine. As Trap Call only uses a fewvectors, you the user have well over a hundred left!Space does not allow me to do justice to, what to mymind, is one of the carefully thought out andrefreshing (no pun) monitors that I have seen for along time. Drawbacks? Well, often programmes aredesigned to run at this address so you'll have somerelocating to do. The other drawback is slightlygreater execution time.

The relative call simulation is used to great effect inthe PIC or Position Independent Code. The Y indexpointer is used in conjunction with this to enable anyprogramme written with PIC to run elsewhere in thesystem. The range is limited to +/- 128 bytes but bywriting in extension routines within this range it ispossible to extend the range to anywhere. ManyRML programmes are written in this way and whenloaded are automatically loaded at the top end of theavailable RAM (stored as HIMEM in COS). HIMEM isthen altered to its new lower value and if desiredanother programme could be loaded ad infinitum(well almost). Each programme then uses PIC to findout where its been put!!

Yet another facility is the Page mode should youwish to address >48K of memory. Of course, you'llneed a Memory Extension Board plus moreRAM/ROM but that's a small price to pay and of littleconsequence. The main point is that RML haveprovided for you.

One cannot end this exhaustive (and exhausting)review without mentioning their 8K Basic which isgood, very good. There are many extra facilities, inparticular the Line Editor, which one would normallyfind in Extended Basics of greater length. This Basicis also very fast - look at the Table in John Coll'sarticle in Vol 1 No. 1 of PCW . It isn't the fastest butits times are still very respectable. How do they doit? Well, for a start all the Basic command words like'Print', 'If' are encoded as each line is entered into aone byte code with bit 7 set. This means that you cansave on memory space, and speeds up programmeexecution time as each programme is part compiled.They are continuing to refine their Basic and, in fact,are up to Basic which has graphics to: -

1) Plot x,y,z where x = 0-79 else errory = 0-47z = 0 = dot off

1= dot on dim2= dot on bright

2) Graph 0 = Normal scroll1 = Clear top 20 lines, scroll bottom

4.2 = No clear, 4 line scroll.

Another facility is to execute 'Control S' beforeexecuting the programme or listing command as thisstops the scroller. Each 'Control A' then scrolls onepage and then stops the scroll and programme.'Control Q' restores to normal operation. It is veryeasy in this mode to wonder why the programme hasstopped (you've forgotten to press CTRL A) and soRML give you a clue by switching the cursor off.

52


Personally, I think it would be better to blink it on andoff. Another minor irritation is the carriage returnautomatically on character 40 and not 41 as oneexpects.

During a visit to RML, I was shown their newAssembler and Text Editor in a working system withtwin cassette machine floppy disc and TTY. Thesoftware also caters for paper tape. Again space andtime did not permit a thorough examination but itlooks as if it is very powerful, very flexible and verygood.

The summing upObviously, many of my criticisms were purelypersonal and should be taken as such. Technically,the product was excellent (the boards really are awork of art!) and there were many innovative andingenious design features. Much though has goneinto the design to make it as flexible and easy tointerface to as possible. The styling is a bit functionaland the general construction could be tidied up. Themonitor ROM and software backup are excellent and

FFFF

C000

8000

4000

3000

User RAM

2000 Assigned

10000E00

Control ports

0800 VDU

COS Monitor

0000

4100 Start of userRAM

40FF Monitor stack

I/O vectors

4000 Monitor variables

3FFF

Unassigned

2000

1FFFCan be used for

1000 ROM

OFFF Control ports,0E00 KBD, User I/O

ODFF Memory mappedVDU

0800

07FF

COS monitor

0000

continually being updated and improved. Existingcustomers can get either free or at cost the fastscroller, fast tape I/O, Intel loader and TTY routinesfrom RML. I just have the slightest unease about thelack of consistency between some of the charactersused for commands and terminators. There is a hintof illogicality about some of them - probably as aresult of this 'fine-tuning' to the system. And don'tforget, you don't need to take out a secondmortgage. You can buy a 280Z version whichconsists of CPU board and VDU board fully built andtested (soak tested for two weeks as are all theirmachines) for £400. If I was on a tight budget, thenthis is the version that I would buy. Otherwise anybusinesses, firms, institutions, schools etcconsidering buying a good all-rounder could do a lotworse than take a long hard look at any of the RMLversions. And don't forget, you aren't necessarilytied to buying any extra memory from ResearchMachines.

To conclude, a few quirks but in general anexcellent machine. I was sorry to see it go.

MEMORY MAP

48k bytes of user RAM possible without

paging

May be used for ROM etc.

256 bytes

512 bytes

13K bytes but physically onlylK

2K bytes

53


F011111EASYPIECES

Neil Harrison

Thou art false but thou art beautiful . . .

Pseudorandom Number Generator

A sequence of random numbers for use in micro-processor games and simulations is usuallygenerated by software. For application wherememory is limited or software generation of randomnumbers is too slow the following circuit can solvethe problem. Strictly speaking the circuit produces apseudo -random (p.r.) sequence of numbersrepeating after 255 shifts. The shift registers (IC1 andIC2) can either be clocked every time the micro-processor reads a random byte or allowed to 'freerun' on the system clock. Presettable shift registersmust be used since a p.r. sequence will not begenerated if the shift registers contain all zeros. Alogic 1 on the RESET line for a minimum of one clockperiod will load non -zero parallel data into theregisters to avoid this problem.

Longer sequences can be generated by increasingthe number of shift registers and altering thefeedback connections. Table 1 gives the feedbackconnections and sequence lengths for p.r. sequencegenerators using up to 16 shift registers.

PSEUDO RANDOM 6 BIT NUMBER

8 BIT PSEUDORANDOM NUMBER GENERATOR

Shift registerstages

feedback bits sequencelength

8 2,3,4,8 2559 5,9 511

10 7,10 102311 9,11 204712 2,10,11,12 409513 1,11,12,13 819114 2,12,13,14 1638315 14,15 3276716 11,13,14,16 65535

Table 1.

Multiplication's the Name of the Game8 by 8 bit Multiply RoutineThis little 12 byte subroutine for the Z80 will multiplytwo unsigned 8 bit numbers together to give a 16 bitresult. The subroutine expects to find the multiplierin the H register and the multiplicand in the Eregister. The result is in the HL register pair.

POWER,

;r6A

+EV 6ND

Plqn1; 000008000000

=00080000000000100 060102 160104 60

CLOCK 0105 290108 300108 190109 100108 C9

PAIR.

= 10008 MPY. LD 0,800 LD D,0

LD L,D

01

MPY1, ADDJR

HL,HLNC,MPY2

ADDFA MPY2. DJNZ

FH4,yDE

RET

Z80 UNSIGNED 8 BY 8 BIT MULTIPLY

ENTER WITH THE TWO 8 BIT NUMBERSIN THE H AND E REGISTERS. THE 16BIT RESULT IS IN THE HL REGISTER

,LOOP COUNTER,D REG = B

,L REG = 0

;SHIFT HL LEFT;SKIP IF C FLAG.0;ADD IN MULTIPLICAND;LOOP UNTIL DONE:RETURN TO CALLER

54


Scrooge's DelightEconomy Potentiometer input for your MPUWhenever your MPU needs to know the position of apotentiometer the usual approach is to use anexpensive Analogue to Digital converter to measuresome related voltage. For applications such as gameswhere stability and accuracy are not too critical thefollowing circuit and program can provide a verycheap alternative.

The circuit, fig 1, uses the good old 555 timer inthe astable mode with the potentiometer R1

controlling the output pulse width. The outputwaveform is shown in fig 2 and the times t1 and t2 arerelated to the component values by the following: -

t1 = 0.685. R3 .t2 = 0.685. R3 . C1 ± 0.685. Ri

The value of potentiometer R1 and therefore itsposition is proportional to (t2 - t1). Resistor R2 is toprotect the 555 and has such a relatively low valuethat it need not be included in the timingcalculations.

The output of the 555 is monitored by a single biton an input port on a Z80, on my system bit 6 of port4 was used.

555

fig. 1

fig. 2 555 Output Waveform

Listing 1 shows a Z80 code subroutine whichreturns a value in the HL register pair, correspondingto the time (t2-t1) and therefore the position of thepotentiometer. Time t1 is subtracted from t2 bystarting with H L=0 and counting down during t1 andup during t2. Before this can happen the subroutinemust make sure that counting starts at the beginningof t1. Since the subroutine finishes its counting at theend of t2 we can use this to indicate the start of a newt1. To do this we use the trick of calling the countingloops starting at POT1 as a subroutine. At the end ofthe next t2 the HL registers are set to zero and thetwo loops starting at POT1 and POT2 perform the(t2-t1) calculation. When the end of the next t2 isreached the routine returns to the main program withthe value in HL corresponding to the position of thepot.

For the component values shown in fig 1 and a Z80running at 2MHz the range of values returned in HL is3 to 500 decimal. This can of course be increased ordecreased by altering the value of C1. The worst casecalculation time taken by the subroutine is 2 x It1+t2max).

00000080

POTENTIOMETER INPUT USING A 555 TIMER

0080 THE OUTPUT OF THE 555 15 CONNECTED TO BIT0000 6 OF INPUT PORT 4

00000008 THE ROUTINE RETURNS WITH A NUMBER IN NL0000 PROPORTIONAL TO THE POSITION OF THE POT.00000080 CD 06 00 POT, CALL POT1 ;WRIT FOR THE END OF T20003 21 00 00 LD HL,0088 ;INITIALISE THE COUNTER8886 20 POT1: DEC HL ;REDUCE COUNT BY 1

0007 DB 84 IN A,(4) ;GET THE PORT INTO A

0089 CB 77 BIT 6,11 ;TEST THE BIT0000 28 F9 JR 2,0071 ;IF 0 WE'RE STILL IN T1

080D 23 POT2 INC HL ;INCREASE COUNT BY 1

000E DB 04 IN R,(4) ;GET THE PORT INTO A

0018 CB 77 BIT 64 ;TEST THE BIT0012 28 F9 JR NZ, POT2 ;IF S WE'RE STILL IN T20014 C9 RET ,RETURN TO CALLER

BONUS TIMEExtra Z80 instructionsMany microprocessors can do things that the manu-facturers either haven't bothered to mention or don'tknow about themselves. The Z80 is no exception tothis and a number of previously undefined opcodescan be used to good effect.

Perhaps the most interesting extra is the ability touse the high and low bytes of the index registers IXand IY as if they are just ordinary 8 bit registers. Thetrick is to notice that opcodes for instructionsinvolving the index registers are the same as forthose using the HL register pair but preceded by DDor FD. For example loading the HL register pair withFFFO is done by the three byte instruction 2100 F8.Loading the IX register with the same number takesthe following four bytes DD 21 00 F8. This alsoapplies to references to H and L as individual 8 bitregisters so if,6A = LD A,L ; Load the accumulator from registerL

then,DD 6A = LD A, IXL; Load the accumulator from thelow byte of the IX register. References to the Lregister will give access to the low byte of the indexregister, the H register will give the high byte. Thistechnique can be used for all Load, arithmetic andlogical operations and probably lots more.

Not quite so useful is the extra shift instructionwhich uses opcodes CB30 to CB37. These codesgive an arithmetic shift left similar to the SLAinstruction. The difference is that bit 0 of the registeris filled with a 1 instead of a 0. Perhaps someone canthink of a use for this!

Thanks to Mike Banahan and Mike Blandfordfor details of these extra instructions.

55


Graham J. Trott

SCRAMBLE is a word game for two or more players.I conceived it after realising that there is an importantelement missing from games played against a

computer - the competitive element, or thesatisfaction of beating a human opponent. After all,a computer is a somewhat emotionless creature - itdoesn't sulk or lose its temper - neither does it gloatwhen it's winning, and in any case there's always thereset button if it gets too smart, isn't there?

Once we decide to write a game for people to playagainst other people, the computer is just there toact as dice thrower and scorekepper; in factSCRAMBLE can be played just as well with a pack ofLEXICON cards and a score pad. For some reason,however, there's more appeal when the TV set is thegame board; even children seem to have no difficultyin concentrating on the game for quite long periods.Besides, our computer isn't earning its keep while wesit around playing cards!

Since VDU's deal with alphanumerics, a wordgame is an obvious choice, and since I had ruled outmost existing games, I had to invent my own. Thisturned out not to be too difficult. Word games are (tome at least) attractive since they depend less on blindluck and more on mental effort. Although the rules ofSCRAMBLE are particularly simple, the program hasto perform a certain amount of gymnastics withstrings. I have drawn an outline flowchart, but the

detailed operation of the functional blocks of theprogram are best understood by studying the listing.The particular features of SCRAMBLE that make iteasy to implement are 1) the display is simple and 2)the only secrecy involved is in the players' intentions.

HOW THE GAME IS PLAYED

The program first asks for the name of the players. Itwill go on asking until a null return is given (just acarriage return). As the program stands, the arraysfor the players' names, the letters they hold and theirscores are not dimensioned; in most versions ofBASIC this is OK but sets an upper limit of nineplayers. In practice, between three and five playersgives the best game, and if your VDU only has a 16 -line display then seven players is the upper limit.Dimensioning all arrays to six or seven elements willsave storage space, as will restricting the number ofelements in a string to 15 or so.

Once all the names have been entered, the maindisplay is set up: -

PLAYER LETTERS VALUE SCORE

TOM 0 0DICK 0 0HARRY 0 0LETTERS AN J*M R 0 KVALUES 1 1 8 0 3 1 1 5

WHAT WILL YOU TAKE, TOM?

The player addressed may do one of three things.Firstly, he (she) may take one of the letter from thepool by typing it followed by carriage return (CR).The star is a joker and can be used as any letter, lateron in the game. The chosen letter appears againstthe player's name, and its value is added to thatalready existing under VALUE. The next player thenhas his turn. As a letter is removed from the poolanother takes its place, so as to maintain a constantnumber (5 more than the number of players). Lettersare randomly chosen for the pool, with the value ofeach letter and its likelihood of occurrence beingspecified by the DATA statements at the start of theprogram. Thus if you feel that the letter D is under-valued and occurs too frequently, the numbersfollowing it can be changed to (say) 3,3.

The second choice open to a player is to form aword with some or all of the letters that he hascollected from the pool on previous turns. To do thisjust type the word, which must have two or moreletters. It's not necessary to type the star; typeinstead the letter it represents. The program checkswhether you do in fact possess the letters requiredfor your word, then asks "ANY OBJECTIONS?". Ananswer beginning with "Y" will cause the repeat ofthe question "WHAT WILL YOU TAKE?", and anyother response (including CR) will tell the program toaccept the word. The letters in the word are removedfrom the player's stock, and his score is calculatedaccording to the following formula: -

NEW SCORE = PRESENT SCORE + (NO. OFLETTERS USED - NO. LEFT BEHIND) * (VALUEUSED - VALUE LEFT BEHIND)

with a minimum score of 1. It is obvious that there isa lot to be gained from having a long word, althoughonly if you use all of your letters. The winner of thegame is the first person to have 200 points, althoughthe program only checks to see if someone has wonafter all players have had their turns.

56


The third action a player may take at his turn, as analternative to either of the above, is to change thepool of letters, in the hope that a crucial letter willappear for his next turn (and not be snapped up byone of the other players). To do this just type CR. Anew pool of letters will then be selected by theprogram, and play will pass to the next person. Thenumber of times a player may change the pool isdependent upon the number of players, but variesfrom two to five, and is set by lines 1700-1750 of theprogram. The only way of getting rid of a letter oncetaken is to use it in a word, so it is unwise to take a Qunless one already holds a U.

START

GET PLAYERSNAMES

SELECT POOLOF LETTERS

SET UPDISPLAY

GET PLAYER'SCHOICE

HOW MANYCHARACTERS

ENTERED?

>1

SELECTNEW POOL

TAKE ALETTER

MAKE AWORD

YES

SET UP FORNEXT PLAYER

THE FUNCTIONAL BLOCKS OF THE PROGRAM

100-130

140

150-190

The letters, with their frequency ofoccurrence and values.Sets N to the total no. of letters (atpresent 83).Control Characters - these willdepend on the terminal used. Thevalues in the listing are for a SWTPCCT -64. (Note: If the interpreter sendsout a line erase character as part of itsCR-LF string, this must be changed toa null. If this is not possible then deleteline 2022 and remove the "A9g1" in line2090.) The purpose of K2 is to deletethe cursor while the display is beingupdated.

1000-1060

1700-1750

2000-2090

3000-3150

3200-3290

4000-4290

5000-5520

7000-7020

SI FMK F II

81000.0012861308148015081.

01.8506051005.85381.1131060

1.0201040.564440

3120

17181720

1;r0lees1810182018,01.820182020

.30.1520402.220502052

.7020002.0256025..203000300538163820.401:g2.02670

3.031003130314031503170318031.1.032.32.22.832053276320032.3-50035.3.52.4000.364650

10704.0409042.00

4,40

42704.3642904.0

5,0050.502050305.05850

5500

:91049207000

7030

Gets the players' names and finds outwho starts. Makes a note of who playslast (X9).Decides how many changes eachplayer is allowed.Sets up the display and asks the firstplayer for his choice.Takes one letter from the pool and putsit against the player's name. Replacesit with another selected at random.Changes the pool of letters for the nextplayer.Removes the player's word from hisstock of letters. Asks for approval thencalculates his new score.Sets up for the next player and checksif there is a winner yet.Subroutine to select a pool of letters.There are 5 more letters than players.

2.0 ABE. F.041. SAS, SOMBUTEABFIRE 8K BASIC

FOR 1.1T027 REA.... V 1..011 NEXT! RESTOREK Y HOW LIP14.1.REM CURSOR ON/OFF

ROMK3421 REM ERASE LINE

ERPSE FRAMEREM SCROLL/PAOE TOGGLE

RE: INTRODUCTICN AND NFRIING FLAYERS

RN FIRST PUT VDU IN .61. MO. HOME CURSOR PRO CL.R SCREEN

PRINT C1409,14.....C.T. P.PRINT "SCRAMBLE THE COMPUTER WORD OFRIE" PRINTPRINT "PLEASE)X YOUR NAME. PRINTPRINT PLOVER INPUTI.CP, IFP40,".T.N105.0

GOT01030

1F..THENX9PIF 3 THENC.5IF P-3 14.144IF P44 THENC1.311c F,c1 IMENC1.2FOR 1.TOP CCII.1.1 NEXT1.INT "EACH PLAYER CAN CHANOE .E LETTERS ,C1."TIMF-SGOSUB 7060REMREM SET UP DISPLAY FWD INPUT CHOICE

TN' HOME CURSOR AND C.. SCREEN . BEFOREREMPRINT HRECK1. *144C140.8"LETTERS,.2>c

PRINT "PLOVER, TA. W., TABC15). TVFLUE SCORE"PRINT FORIR1TORIF AS. TMENIFIC,X1 1.6NPRINT 00T02635PRINT P4,11.PRINT TABC26, LE, I, TABC361,, I Tc TAW.). SC I c

NEXT I PRINTPRINT ETTERS,FOR 41L.LENCL., P010.11.0.1.1),.NEXT[. PRINTPRINT .ALUESc. TAW.,FOR 1.TOLENCVS,.NEXT[. PRINTPRINTPRINT PRINT.WHAT WILL VOU TAKE. ,PClOcPRINT .R.1(3). INPUT..41 PRINTCHR4(14,11.REMREM TF.E ONE LETTERREH

THEN2266IF LENCA1011TFEN4006FOR 1411.9IF MI.CL. 1...R.THEN3050

PRINT CH.C13......R....11,, cc I I, 00102070L000400041.

,JS.T.SCX)41.11.C., I...EFTCY3.1-11.1014TS(.1.09-12N1F110,14.4.D.TT

RESTORE H2V N1.1-1.12.1.120THEN31396

RESTORE 0.9111. VS.H4STROCV)VCX.0 FOR141TOLENCV.X,>VCRIRY.C.1.01CMI., YE, 1.1C.).NENTIGOTO 5006REMREM CHANGE THE LETTERSpF,COO.0 <10 -I IFCCXT/OTHENOCKUB71030 GOT03.250

0010 3.0F.INT CHR441,,,CHR.11,01114,112.C..<11,..RSC11).

XF1.IFX.4.1TNEMX41IF 10.4NEWRINICHE60,21.1 00T02040FOR 141T.. IFS" /GS. THENS.SC 1 J. 19-1NEXT 1 FS9.00 TIENPRINTCHRECK.c .00T02.0PRINT PRINT PRINT PRINT 00T05630RE,0. H

REMIF LENCLiCX".THEN4098S1..S4.4...00.T.V.(X).FOR141TOLENCI14>

IF M104 C T THF,4200NEXT J.FORJ41TOLEINS.IF J. 1).. TWEN2110.XT JPRINT GOT02046

.EFTRCS4 stoasisS14.1,14J.1, TTERLEFTRCT. J -1).1011T., TS. LENTEJ-a)

TXTXIO

NEXT INPU.. CAJECT1.S, IRIF LEFT11114.12-", IHELYRINTc.*U.,,

si-si-Y00 1E51,1 THENS1F1SRFLENCFMJ-LENCLS<X>1. 1E5..1 THENE2R1

PRINT CHREC1...1.001).REN

:1;1, SET UP FOR .KT PLAYER

IF X4,X9THEN5568FOR 1.1TOP IFS< li,S9 T.NS.S<I). 19.1NEXT I IFS942. 1.05500PRINT .444 ',P.M." IS THE WIN., *RFTPRINT CHR4.1..) PR1NT.SCORES -..PRINTFOR 141TOP PRINTPEC IL TREN.26,SCI., RENTZPRINT PRINT PRINT STOPRIF,. %RR'S rICENX41PRINT C.4,13,...4.4.T.FOR 11TOP. FIRINT.R4,11). NEXT]GOTO 2020REIT

REM SUBROUTINE TO SELECT THE POOL OF LETTERSREM

READ ....V N1.1 N2 1.1.TNEN7010VI A.181.1

The program requires a BASIC with full stringhandling facilities, including string arrays, and willrun in 4-5k (depending on the number of players)over and above the interpreter. Where multiplestatements on one line follow an IF . . . THEN it is

essential that on the test failing, control is passed tothe next line, not to the next statement on the sameline. This may seem obvious, but not all BASICsfollow the same conventions.

57


COMPUTERin THECIASSROOMTEACHInGPACKAGESCharles Sweeten

It is becoming apparent that within the next fiveyears many schools will have the opportunity to havea computer in the classroom. Those that are luckyenough to have one now will realise that there arepossible ways of using the computer other than forteaching computing. One such way is to use thecomputer as a tool to illustrate a method or aproblem to a class.

You need a computer that will run a fast andefficient version of BASIC. The BASIC must handleprogram files easily and quickly, and ideally theseshould be contained on a disk. You will need a largescreen which can be connected to the keyboard sothat a whole class can read it at a time.

The teaching package consists of teacher notes,pupil notes, details of knowledge required, togetherwith a program that requires no knowledge ofcomputers and virtually no knowledge ofprogramming. Knowledge of how to control a

terminal is necessary of course. The program itselfshould be foolproof and this is no easy matter toachieve. The program should be self explanatory inoperation and it should not be necessary to list it ormodify it in any way once it has been adapted for aparticular system.

The idea is that the teacher should lead the classtowards the point to be illustrated, and shouldprovide the theoretical or practical experience thatmay be required. Having achieved a sufficient level ofcompetence in the class the package may be run inthe way that is best suited to that package. It is ofcourse essential that the requirements of knowledgeto be expected from the class are known exactly.Otherwise a class will be as unhappy as they usuallyare with a chapter of a book that is either too hard ortoo easy for them.

I will now take these points together with anexample from A -level Mathematics. Many singlesubject mathematics courses include some statistics,and where this is not the case there is usually aseparate Statistics A -level. One of the most difficultconcepts to grasp is that of taking samples and ofmaking judgements based on such samples. Mostteachers would agree that the best way ofapproaching this topic is by experimental work, butthere is frequently little time to do much that isworthwhile. It would be pleasant to get away fromthose boring experiments with dice, and measuresomething which has relevance to the class.Different classes will have different interests, butmany who study statistics will have an interest inbiology or social implications. The experiment that I

have designed is about a river that is being pollutedby a factory.

The class is asked to take samples from thepolluted river in order to establish whether it issufficiently polluted for someone to take action ofsome kind. For example, if fish are dying then theRiver Authority may take legal proceedings. They aretold a little about the difficulties of sampling the riveradequately, and they are told about the difficulty ofestimating the accuracy of their measurements. Theyhave to make a decision as to how the results are tobe tabulated in groups and they have to decide howmany measurements to take to establish any sort ofan answer.

The knowledge required by the pupils isexperience in using the Normal distribution, togetherwith some knowledge of the relation between themean and standard deviation of the population andthose of the sample. In this case the mean of thepopulation is the true pollution level of the river(whatever that may mean), and it is this level that weare attempting to discover. Some discussion of theactual process of sampling will be needed as mostpupils have little idea of what this involves. In thiscase discussion might centre on making up a plan fortaking the samples, in terms of point of the river,depth, time of day, and etc . . It is also worthdiscussing how long it would take to carry out thework (don't forget travel and meal breaks).

The class should have discussed the effect ofarranging measurements into groups and should beaware that choosing the group widths too wide ortoo narrow will conceal the shape of a distribution,but the package is designed to let them see this forthemselves.

Pupils should also have some knowledge of thedistribution of sample means so that discussion cantake place over the range within which the truepollution level may be expected to lie.

In this particular package teacher and pupil notesare the same but there is no pressing reason that theyshould always be so. The notes are shown in figure 1.

Figure 2 shows a sample run. The user first decideswhether instructions are wanted. They then decidehow many samples they want to take from the riverand in return they are given the time it will take to dothis and process the results. (It is worth discussingwith a class the effects of too few or too manysamples).

They are then asked to decide how to classify thereadings in groups. This is a practical point that isoften glossed over because of the lack of realexperimental work, but makes a great deal ofdifference to the usefulness of statistical work. If theuser requests a ridiculous number of groups theprogram will refuse to cooperate and asks for morerealistic instructions. The program will take samplesfrom a Normal distribution of pollution levels, whosepopulation mean and standard deviation arethemselves chosen randomly. The full statisticalcalculations are then set out. The package may bemodified and left on the computer system for pupils,and in this mode all that may need to be done isdelete lines to , so that the pupils must work outthe figures for themselves.

The package has been tested thoroughly on a classof 2nd year sixth form boys at Oundle School. Ofthese perhaps half may fail A -level Mathematics inthe summer and the rest are not expecting highgrades. They lack confidence and have littleenthusiasm for anything that will further diminishthat confidence. They know very little aboutcomputing but have been quite enthusiastic aboutusing the package on their own. I have often found a

58


couple of them at a terminal trying out the packagesthey have done some four months earlier. It is alwayshard to tell whether some educational methods arean advance on others, but I think the packageapproach has given this particular group moremotivation than they would have had otherwise.

MATHS RIVER STATISTICS Normal: Esti-natal./

You will be supplied with information taken from apolluted river. The pollution is caused by a factorywho will only make an effort to clean up their effluentwhen enough pressure is put on them. You are beingasked to provide the evidence. The effect of actionagainst the factory is unfortunately only temporary,and afterwards pollution levels gradually return to thesame levels as existed before. The factory neverreleases information about any action they may havetaken; nor do they admit that they have taken action.So it is impossible to judge when to start bringingpressure to bear other than by sampling the river.

Pollution levels appear to vary throughout eachday and vary at different depths and places, but nopattern has been observed. The method ofmeasurement gives variable results in any case, dueto causes such as the time between the sample beingtaken and being analysed, the temperature of thewater and of the air, etc. It is not possible to doanything about this as sufficient funds are notavailable.

You are required to show that the samples takenfrom the river form roughly a Normal distribution. It isknown that the level at which fish begin to die isroughly 15 pu's (pollution units) and you are asked toprovide evidence as to whether immediate action isnecessary or preliminary pressure or no action.

The machine which analyses the samples givesreadings correct to .0001 pu, but this is no indicationof the actual accuracy of your samples. In fact thelast inspector who did the work thought that hissample measurements were only accurate to 1 pudue to the various factors such as temperature,depth, etc. that we have already mentioned. Youmust decide for yourself what accuracy you want inthe table of results.

The readings will be grouped by the technicianswho do the analysis into groups and they will need tobe told the width of these groups.

You must decide before you start how manysamples you think will be needed to establish yourresults. Remember though that sampling takes time.The width of your groups will depend on the numberof samples you take; the fewer samples the widerwill be your groups. The idea is to get a histogramwhich is more or less smooth.

You will first be asked the number of samples : youmay answer any number from 1 to 10000.

The group width should be anything from .0001 to10.

Here is a sample of the program :

You must decide before you start how many samples you think will

be needed to establish your results. Remember though that sampling

takes time. The width of your groups will depend on the number of

samples you take; the fewer samples the wider will be your groups.

The idea is to get a histogram which is more or less smooth.

You will first be asked the number of samples : you may answer

any number from 1 to 10000.

The group width should be anything from .0001 to 10.

Here is a sample of the program :

HOW MANY SAMPLES ARE YOU TAKING FROM THE RIVER 7 200

IT WILL TAKE 48 HOURS TO PROCESS THE RESULTS.IS THAT ALRIGHT 7 YES

YOU ARE GOING TO GET READINGS SUCH AS 14.8647THE READINGS WILL BE GROUPED IN A FREQUENCY TABLEHOW WIDE DO Y0'.: WANT EACH GROUP TO BE ? .5

FREQUENCY TABLE

11.51212.51313.51414.51515.51616.51717.51818.5

- 12 XXXXX 3

- 12.5 XXXXX 3

- 13 XXXXXXXXXXXX 8

- 13.5 XXXXXXXXX 6

- 14 XXXXXXXXXXXXXXXXXXXXX 14

- 14.5 XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX 30

- 15 XXXXXXXXXXXXXXXXXXXXXXXXXXXXXX 20

- 15.5 XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX 33

- 16 XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX 27

- 16.5 XXXXXXXXXXXXXXXXXXXXXXXXXXXXXX 20

- 17 XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX 22

- 17.5 XXXXXXXXXXX 7

- 18 XXXXXXXX 5

- 18.5 XXX 2

- 19 0

PRESS RETURN WHEN YOU HAVE READ THIS

MEAN . 15.1825 PU STANDARD DEVIATION e 1.33574 PU

ESTIMATE OF POPULATION STANDARD DEVIATION . 1.3391 PU

STANDARD DEVIATION OF THE MEAN = 9.46883E-02 PU

95% CONFIDENCE LIMITS FOR THE MEAN ARE14.9969 PU TO 15.3681 FU

0010 REM ..... RIVER.*"..0020 REM THIS VERSION PUNE IN ZPL 8: EAST:.

0030 REM TESTED ON RN 3412.0040 CLEAR 2000050 RANDOEIZE0060 PRINT CHRS(12)0070 PRINT0080 REM SCREEN SHOULD NOW IL CLEAR0090 FAINT0100 PRINT "SIMULATION or RIVER POLLUTION"0110 PRINT 0120 PRINT0130 PRINT0140 PRINT 'DO YOU WANT INSTRUCTIONS0150 INPUT Qt0160 IF LEFIX(08,1)..Y" THEN SOTO 11000170 PRINT0180 PRINT0190 LET 0.14.4.RND(9)0200 LET S..50ND(9)0210 PRINT "HOW MANY SAMPLES ARE YOU TAKING FROM THE RIVER ";

0220 INPUT N0230 LET DmINT(N/55.1)0240 PRINT "IT WILL TAXE .10250 IF 0.1 THEN PRINT "24 HOURS.;0260 IF 0.2 THEN PRINT .48 HOURS";0270 Ir D>2 THEN PRINT DODAYS.;0280 PRINT TO PRODUCE THE RESULTS ."0290 PRINT " IS THAT ALRIGHT ";

0300 INPUT At0310 IF LEFTS(A8.1)."N" THEN GOTO 01700320 PRINT0330 PRINT "YOU ARE GOING TO GET READINGS SUCH AS 14.8647.0340 PRINT "THE READINGS WILL BE GROUPED IN A FREQUENCY TABLE.0350 PRINT ^ HOW WIDE DO YOU WANT EACH GROUP TO BE ";0360 INPUT G0370 IF GT(0 THEN GOTO 12000380 LET .10.711AT WILL GIVE MORE THAN 250 GROUPS I"0390 IF 6"S/G>250 THEN PRINT At0400 IF 6.S/G4250 THEN GOTO 03500410 LET RI.M-3"S0420 LET R1.INT(R1/0).G0430 DIM %(6.5/G)0440 LET F.00450 FOR N1.1 TO N0460 GOSUB 20000470 LET Y.INT(10-R11/1)210480 LET X(Y).X(Y)"10490 IF %(Y)>F THEN LET F.X(Y)0500 NEXT NI0510 LET K.10510 REM THE DISPLAY MAY BE WIDENED BY CHANGING 19 TO 400530 IF F4I9 THEN LET 1.19/F0540 REM THIS SECTION SETS A DELAY PROPORTIONAL0550 REM TO THE DAYS REQUIRED0560 PRINT "SAMPLING HAS NOW STARTED"0570 FOR T.1 TO 30000.D0550 NEXT T0590 PRINT0600 PRINT "FREQUENCY DISTRIBUTION.0610 PRINT0620 LET 111.00630 LET 51.00640 FOR 0.1 TO 6"5/6"20650 PRINT RI.(Y-1/4G;0660 PRINT TA8(6)1.- .;0670 PRINT RI.Y.GT0680 PRINT TABOOS)0690 FOR C.1 TO INT(X(Y).1(4.5)0700 PRINT "X";0710 NEXT C0720 PRINT "WY)0730 LET X.R.(y-.5).G0740 LET Ml.MITX(Y)*X/N0750 LET SI.S1.X(Y)AX"X/N0760 NEXT Y0770 PRINT "PRESS An KEY FOLLOWED BY RETURN"0780 PRINT "WHEN YOU HAVE READ THIS";0760 INPUT Qt0500 LET Sl.SQR(SI-ME.ME)0810 PRINT 0)

0820 PRINT ...0110 PU.,.STANDARD DEVIATION ..;0830 PRINT SW PU.0840 LET 52.1.SQR(N/(N-1))0850 PRINT CHRS(10)0ESTIMATE OF POPULATION STANDARD.;0860 PRINT ^ DEVIATION ..:520 PU.0870 PRINT CHRS(10);"STANDARD DEVIATION OF THE MEAN .";0880 PRINT ST/SQR(N);" PU"0890 PRINT CHRt(10)096% CONFIDENCE LIMITS FOR THE MEAN ARE"0900 PRINT 111-1.9652/SQR(N);" PU TO .011.1.96.12/SQRINIO PU"0910 LET M.00920 LET 5.00930 END1200 REM "50 THAT THE INSTRUCTIONS CAN BE READ ON YOUR1210 REM "SCREEN COUNT THE NUMBER OF LINES DISPLAYED AT1220 REM "A TIME --SUPPOSE THIS IS N LINES --THEN INSERT1230 REM "THE FOLLOWING INSTRUCTIONS AFTER EVERY N LINES1740 REM PRINT "PRESS An KEY FOLLOWED BY RETURN"MO REM PRINT "WHEN YOU HAVE READ THIS.;1260 REM INPUT At1270 PRINT CHR8(12)1280 PRINT1290 PRINT YOU WILL BE SUPPLIED WITH INFORMATION TAKEN mom.1300 PRINT POLLUTED RIVER. THE POLLUTION IS CAUSED BY A.1310 PRINT "FACTORY WHO WILL MAKE AN EFFORT TO CLEAN UP THEIR.1320 PRINT "EFFLUENT ONLY IF YOU CAN PROVIDE THE EVIDENCE."1330 PRINT1340 PRINT POLLUTION LEVELS APPEAR TO VARY THROUGHOUT EACH"1350 PRINT "DAY AND VARY AT DIFFERENT DEPTHS AND PLACES, BUT"

1360 PRINT "NO PATTERN HAS BEEN NOTICED."1370 PRINT1380 PRINT YOU ARE REQUIRED TO SHOW THAT THE SAMPLES TAKEN"1300 PRINT "FROM THE RIVER FORM ROUGHLY A NORMAL DISTRIBUTION"1390 PRINT1400 PRINT " IT IS KNOWN THAT TIsH BEGIN TO DIE AT ABOUT 15.1410 PRINT "PI (POLLUTION UNITS) AND YOU ARE REQUIRED TO.1420 PRINT ^PROVIDE EVIDENCE AS TO WHETHER IMMEDIATE ACTION"1430 PRINT "IS NECESSARY."1440 PRINT1450 PRINT THE MACHINE WHICH ANALYSES THE SAMPLES GIVES"1460 PRINT "READINGS CORRECT TO .0001 PU BUT THIS IS NO"1470 PRINT "INDICATION OF THE ACTUAL ACCURACY. IN FACT THE"1480 PRINT "LAST INSPECTOR WHO DID THE WORK THOUGHT THAT HIS"1490 PRINT 'SAMPLE MEASUREMENTS WIRE ONLY ACCURATE TO 1 PU DUE'1500 PRINT .TO THE VARIOUS FACTORS SUCH AS TEMPERATURE, AND.1510 PRINT "EXPOSURE TO AIR, AND ETC.. YOU MUST DECIDE row.1520 PRINT "YOURSELF WHAT ACCURACY TO TAKE.1530 PRINT1540 PRINT THE READINGS WILL IL GROUPED BY THE TECHNICIANS"1550 PRINT "'WHO DO THE ANALYSIS INTO BANDS AND THEY WILL NEED"1560 PRINT.TO KNOW TEE WIDTHS OF THESE GROUPS BEFORE THEY"1570 PRINT "START..1580 PRINT1590 PRINT .....THE NUMB CF or SAMPLES: ANYTHING FROM 1 TO 10000.1600 PRINT THAT ALRIGHT: YES OR NC OR Y OR1610 PRINT .....HOW WIDE IS LACE GROUP: FRCM .0001 TO 10.1620 PRINT1630 GOTO 01102000 FEB THIS GIVES SAMPLES FROM A NOMA.. .511213,710N2010 ELI 3.637.1,(91-32070 LIT T.1247(9)2030 IF 1,..03.111-333/2)/2.50663 THEN COTO 20002000 LEI 1.M.10)5

RETORT:

59


INTERAMputs its cardson the table . . .

THE KEYED -UP 8080This board is a combination 8080A CPU

Er Intelligent Front -Panel. The 12 -Keygold contact keyboard works in conjunc-tion with iK of on -board ROM Er RAM toallow you to examine or deposit data intoany Register, Register Pair, Memorylocation or I/O Port. Single step orSLOW -STEP at a programmable ratethrough your program. The ten 7 -

segment LED's offer a concise OCTALdisplay which all in all add up to controlover your computer so complete itsurpasses any mini -computer! £225

THE SPEAKEASY I/O BOARDSlot this single board into your S-100

BUS computer and you're ready to plugin 3 audio cassette units, a Teletype orRS232 terminal and any 8 -bit paralleldevice . . . graphics display, line printeror paper tape reader. On -board ROM ErRAM pre-programmed with COPEhandles dozens of interfacing and datahandling tasks. COPE can even bootstrapyour programs. £135

THE WUNDERBUSSThis the heart of every S-100 BUS

computer - the BUS itself. No otherbussboard, though, has each busslineisolated by an interlaced system ofground lines and cross -coupled groundplanes plus active termination to preventcross talk and eliminate signalreflection. £75

Wonderful things have happened tocomputers since the introduction of the S-100BUS in 1975. Reliability has shot up like arocket and prices have dropped like a rock.Today there are now over 100 manufacturersmaking plug -compatible computers, circuit -boards and peripherals. INTERAM supply afully comprehensive range of computerMemory boards, Intelligent Front -Panels,Buss -Boards, I/O Interfaces, Control Boards,Extender Cards, Speech-Recognisers and Syn-thesisers. All of which plug directly into yourAltair/EQUINOX/IMSAI etc. S-100 Buscomputer.

ECONORAM II 8K STATIC RAMThis S-100 compatible memory board

features two independent addressable 4Kblocks. All lines fully buffered, tri- stateoutputs, wait state and write strobe selectswitches. Board access time of 450ns.Can handle DMA devices. £135

Al CYBERNETICS SPEECHSYNTHESISER

This is a revolutionary form ofperipheral for your S-100 Bus computer.Interactive computing will never be thesame again once your computer hasactually spoken to you. The unit is a hard-wired analog of the human vocal tract. Allof the information necessary to produceEnglish speech has been programmedinto ROM's which reside on thesynthesiser board. £275

HEURISTICS SPEECHLAB SPEECHRECOGNISER

Now you can have low-cost voice entryto your computer. All models in theSPEECHLAB range plug directly into yourS-100 BUS computer. Use it for scientificand engineering experiment data entryand countless other applications. From£145

ALL THE ABOVE BOARDS ARE FULLY ASSEMBLED,TESTED, AND WARRANTED FOR 1 YEAR

ALL PRICES EXCLUSIVE OF VAT

DEALER ENQUIRIES INVITED

MULLEN OPTO-ISOLATOR /RELAYCONTROL BOARD

Although a computer may performinteresting operations and programs, itsusefulness in controlling real worlddevices - such as motors, audio signallines, display lights, etc. - is severelylimited. This board controls low -powerdevices directly, or triggers heavy dutyrelays or triacs. Additionally, 8 opto-isolators accept an 8 -bit word from theoutside world and send it to the computerfor handshaking or further controlpurposes. £105

MULLEN EXTENDER CARDWhen many boards are loaded side by

side in a computer, trouble -shootingbecomes almost impossible. This usefulaccessory "extends" a given board overthe tops of other boards for easy trouble-shooting and examination. £39

INTERAMCOMPUTER SYSTEMS LTD.

59 MORETON ST.,VICTORIA,LONDON SW1V 2NY

TEL: 01-834 0261/2733

60


Thecomputeris justwhat thedoctorhasn'torderedRobert Johnson

Why is medical data processing still in its infancy? All other service industries have embracedcomputer techniques to the unprecedented benefit of themselves, and their clients. Only the healthservices lag behind: why has so little advantage been taken of data processing techniques?

The problem is not a small one. The National Health Service costs £7 billion a year to run. It hasmore employees than any other service industry, and shortly lin 1979) the drugs prescribed by generalpractitioners will cost £1 billion a year.

Cost, however, is the least of the hazards. The veryfoundation of health care is now in jeopardy, merelythrough faulty data processing. Tragedies such asThalidomide and Eraldin will recur with increasingfrequency simply because the drug informationcannot humanly be processed in time.

Information pollution?These assertions may seem alarmist, perhapssensational, but we do not require a medicaleducation to appreciate their inevitability. Indeed, itmight almost be said that a medical training makes itmore difficult to visualise the simple logistics of thesituation. The quantity of drug information isconstantly growing: new drugs are being marketed,and new facts about old drugs being published.Using conventional methods, this information isdelivered to every GP in the country in the form ofmonthly and annual books with official governmentapproval, if not statutory backing.

Disregarding all the promotional literature andmedical periodicals supported by advertising, eachGP receives a mammoth total of 4 million wordsevery 12 months. This figure could rise in the 1980sto four and then five million.

To computer personnel, two and a quarter millionwords may translate roughly to 10 Megabytes: smallbeer compared with the hundreds of megabytestorages now available. But for the humble GP, theseare not on electronic data files: they are printedwords on paper pages - side by side they constitutea book 6 in. thick; and when one year finishes, thereis another 6 in. for the next.

If this is not deterrent enough, the cards on whichpatients' clinical records are stored are similarlyexploding in size. It stands to reason that as moreelaborate clinical tests are devised, and morebiochemical readings are fished from the bloodstream and elsewhere, the results must be storedsomewhere.

61


And, since the GP sees the patient three or fourtimes as often as do the hospitals, for the client tobenefit from these exertions at all the results must beavailable in the GP's surgery. On average it has beenestimated that each patient's record will grow in thismanner by 8% annually, doubling therefore everydecade.

Picture then the GP sitting at his surgery desk with6 in. of drug books on one side, and a 1 in. wadge ofrecords for each patient on the other. He has toextract relevant information from each 'data bank',compare it with the details being presented in thecurrent interview, and then prescribe an appropriate,economic, safe and effective treatment.

The picture is not quite complete. In a 12 -monthperiod in general practice, I saw 11,365 patients -roughly 200 a week. With a nominal 40 hour week,this allowed a gross average of 12 minutes for eachpatient contact.

If an additional five minutes were required to checkan item that had slipped the doctor's (or patient's)memory, the penalty incurred would be an additional16 hours per week. So yet another natural resource,time, is running out.

One might have thought there would have beensomething of an outcry from doctors, or a little moreactivity on the part of the Department of Health andSocial Security to remedy a situation that is alreadyfar from healthy.

The gulf between need and actionIf the need is there, though not officiallyacknowledged, why has so little been done? Theprimary obstacle is obviously the DHSS. What otherservice organisation could justify spending less than0.5% of its budget on research into improved dataprocessing?

But the chief difficulty is lack of imagination andfear of innovation. It is illusory to hope that modernelectronic techniques can be grafted onto centuries -old medical practices, without profound and radicalanalysis of both.

The illusion persists, and will prove increasinglycostly. Doctors have been reluctant to tackle thebasis and essentials of computation, and fewcomputer experts would volunteer to spend a decadelearning medical techniques from the inside.

Computers accordingly have a bad name in themedical profession, and their progress is thereforeresisted. This is partly because of a few early butsensational failures, partly because of cost, andpartly because there is no real understanding of howtoday's technology can assist so skilled an area asmedical practice.

In my view the practice of medicine will not survivewithout computer assistance. There is an enormousvirtually untapped market for both hardware andsoftware; and there are simple reasons why mostearly medical computing attempts were doomed.

Taking each of these points in reverse order: mostextant medical computing systems store records infree (that is, uncoded) text. The clinical data going inmay be of the highest quality, but there is simply nocost effective method of getting it out, unless one isprepared to forgo all the benefits of data processing.

Medical data must therefore be entered in numericor digital form - a requirement which is greeted withthe profoundest suspicion by the medical profession.You might as well be asking them to write their notesblindfold. Nevertheless 'free text in, free text out'achieves little, at great expense - something whichwas not applauded by a recent Parliamentaryaccounts committee.

Cost is a convenient hedge behind which mostdoctors will hide - a position rendered untenable ofcourse, by the microprocessor. For £1 ,000, a kitusing audiocassettes, domestic TV screens, andperhaps 16K RAM of memory would provide thebasic requirements for a sophisticated medicalcomputing system, to which more could be added indue time.

The numeric codes need not be memorised, butcan be read from the screen by the doctor, or, if needbe, his receptionist or secretary. Being exclusivelynumeric, the data can be tightly packed. Thus theentire 12 months records for a busy generalpractitioner can be stored on a single C60 audio-cassette, costing say 20p in bulk. Five thousandpounds would then cover the entire annual medicalrecords of all 25,000 GPs in the NHS.

Once entered the data could at last be manipulatedelectronically. The machine would cease to be anover -grown electronic notebook, but become anactive processor of the numerically encoded medicaldata. So having coaxed the medical profession intoadopting more up-to-date recording techniques,what can computerisation offer them?

The pay-offFrom a commercial point of view, the administrationof general practice is a gold mine for microcomputer -based management systems. The GP of today isincreasingly paid on an 'item of service' basis: that isto say, of the myriad individual 'treatments' hedelivers, for some he receives a fee which goes upand down depending on fairly arbitrary rules: and forothers reimbursement, which again wanders over awide range. In such a topsy-turvy world, the GP caneasily lose up to £1,000 a year, in 'lost' fees, eitherclaimed too late or not at all.

Leasing a system for f5-£10 a week would be veryattractive to the average group practice, and couldsoon be shown to pay for itself, by running electronicrings around the paper -bound log -jams of the localadministrators. A game of profit and skill that manyGPs would relish.

Prevention is better than cureCost in terms of cash can no longer therefore be usedas a barrier to further medical computing. Costs interms of drug -induced diseases are only just beingreckoned. Thalidomide is well known, and had anefficient drug data distribution service been inoperation at the time, there would have been novictims in this country. Eraldin is perhaps less wellknown, but is even more relevant.

Eraldin is an excellent drug for angina, but it tookfive years to find that it sent some patients blind, andgave others a lethal form of peritonitis which had notbeen met with before.

If such poisonous effects occur at a rate of onlyone in 10,000 or even one in 1,000 the individual GPprescribing the drug has just no chance of pickingthem out from the 30,000 or so symptoms he seesevery year.

Only by pooling data from many if not all GPs cana feasible strategy be adopted. Official proposals formore paper work for newly -released drugs are likelyto sink under their own weight and cost. Meanwhile,we, as a community, continue to incubate other'Eraldin-type' disasters.

Using the system mentioned, every GP (or if he isunwilling, his receptionist) would record thesymptoms presented, the diagnosis derived, and thetreatment prescribed. These would be stored initiallyon a 2K RAM scratch pad, to be read off at leisureafter the surgery onto the 20 pence C60 cassette. 2Kof eight -bit bytes will cover 50 consultations on

62


average, enough to handle all but the most Herc-ulean workload.

Computerisation thus offers the GP savings in hispractice costs, it offers the first practicable means ofdetecting totally unforeseeable side -effects from alldrugs prescribed, whether newly -released or oflongstanding 'respectability'.

I estimate that there are around 500 millionsymptoms presented to GPs every year in the UK, 50million of these are pains, 15 million headaches.Prompt main-frame analysis of data of thismagnitude and topicality would be of unprecedentedvalue, not only to drug companies with newproducts, but but also to the Committee on theSafety of Medicines, and of course to the GPhimself.

It could inform him in retrospect as to theeffectiveness of his particular treatment in delayingor abolishing specific symptoms, or any otherresearch topic which took his interest.

The patient (that is to say ourselves) will benefitmost, when the system is developed to operate inreal time, interactively. Costs would double to say£2,000 capital, £10-15 lease per week, since a floppyor other disk would be required.

Filling in the prescriptionAn example of a computerised system for GPs is theAPRIL system. With such a system, the benefits ofcomputerisation have an impact everyone, especiallythe public, can appreciate.

Indeed, it may well be the case that only publicpressure will induce the DHSS and the NHS itadministers to take a more active interest in suchsystems.

The APRIL system analyses each individual patientrecord, and compares it electronically and infalliblywith the data on the drug about to be prescribed,with the consequence that no known harmful drugswill be received. Drug treatment is optimised to suiteach individual patient.

The basis of the APRIL system is as follows: oncethe clinical data and the drug information have beenencoded in the same numerical code, adverseinteractions between the two may be detectedsimply by subtraction.

Programmes for this are still being developed, butthey need not exceed 8 or 16K. Compared with theMegabytes of lexicons, and the large data banksrequired for the 'free text' exponents, the amount ofhardware required is minuscule.

One further technical point. In the currentprototype, only 4K bytes are set aside for the 10 -letter words used for visual decoding of the numericson the screen. This size of memory is adequate forthe simplest system which covers those items whichoccur with greater frequency than once weekly onaverage.

The law of diminishing returns applies to theremainder, which could with this basic system onlybe covered in generalities. Larger studies couldreadily determine which additional items needincluding. Similarly, it is possible to envisage aheuristic model, whereby the drug names used mostcommonly by a given doctor receive more space inhis personalised system.

With the APRIL system operational, it would beimpossible for you or me to be prescribed a drug thathad previously been known to poison someone withyour (or my) particular clinical history. Drugssuspected, for example, of affecting embryos (aswith Thalidomide) would automatically be screenedagainst possible indications of pregnancy on the partof the patient, BEFORE the drug could be prescribed,

so providing a degree of security in safe prescribingquite unobtainable elsewhere or earlier.

Similarly, a past history of what could beconstrued as heart disease would automatically, andwithout further fallible human intervention, exclude adrug or combination of drugs found (anywhere in theworld) to have been detrimental.

There are many points which cannot be includedhere: drug actions and interactions are immenselycomplex. Some drug interaction patterns requirearrays of up to five or even 10 dimensions - acomputation somewhat beyond the standardcerebral capacity of the normal hard-working GP.

The paper epidemicA cautionary tale to end with. Drug -induced diseaseis an epidemic arising from the growth of new andmore powerful drugs. Its cause is inefficient dataprocessing. It is therefore analogous to the earlierself-inflicted epidemics, such as typhoid and cholera,which are consequences of inefficient sewageprocessing.

A combination of public and professional effort isneeded to contain both types of epidemic. Onefamous episode from the annals of public healthconcerns the doctor who averted just such anoutbreak by removing the handle of the 'BroadStreet Pump'. He had deduced that the water from itwas infectious, and acted accordingly.

In our present case, however, the pump handle inquestion is the actual pen with which the doctorwrites and suggestions for improvement may not,as I can testify from experience, be immediatelywelcome.

I should not be at all surprised to find patientsdeploying computer systems before their doctors do.After all, as actual consumers of the treatment, theirinterest in the safety of the prescription is that bitkeener.

63


PCWThe

OPENAmateur

PAGE

Computer ClubView Mike Lord Homer G Homrew

So PET has finally arrived in the U.K. - or to beprecise a system has been sighted and CBM are nowwilling to take your money and put your name ontothe queue. A pity that they are offering only thelarger, 8K, version: £695 is a bit much for the averageamateur. However, CBM have obviously decided toaim for the top end of the market while demandexceeds supply, and why not? This is a fascinatingmachine, the graphics look superb and theirextended version of BASIC seems to be powerfuland easy to use. A pity they didn't go for a more solidlooking keyboard than the calculator style 'FingerFumbler' that they have fitted.

Nearer to the size of the average amateur's wallet;the saga of Lynx's NASCOM 1 continues. At thetime of writing I haven't heard from anyone who hasactually received a kit; but Lynx say that the PCBproduction problems are now ironed out and within aweek or so deliveries will start. Although not a Z80enthusiast myself (it's too complicated a device forme to remember how to use it from one weekend tothe next) I think NASCOM is a significant milestonein the U.K. amateur computing scene, and lookforward to hearing users' experiences with it.

Both PET and NASCOM 1 attracted greatattention at a recent ACC London Group meeting,which also featured several homebrew systems,including Neil Harrison's Z80 running some nicegame programs; where does he get the time from?

The Midlands Section of the ACC has beenholding successful meetings at approximately six -week intervals, anyone interested in attending shouldcontact Ray Diamond, 27 Loweswater Rd.,Coventry CV3 2HJ (telephone Coventry 454061) forthe latest schedule.

One group which obviously believes in planningahead is ACC North West. They are running a seriesof elementary meetings, which occupy the first half-hour of each meeting and are followed by a moreadvanced talk, starting at 6.30 in the NationalComputing Centre, Oxford St., Manchester. Thecurrent program is;

4 May Computer Arithmetic8 June Computer Architecture

However, in case of any last minute snags, anyonewishing to attend should confirm with Ken Horton,50 Lymfield Drive, Worsley (061 799 0192).

Mancunian Micro Mashers: the Northern RadioSocieties Association's exhibition on Sunday 2nd

64

April featured microprocessors and personalcomputing. Exhibitors booked included ComputerWorkshop, Comart, Lynx, Newbear and LPEnterprises. .

Bob Cottis is starting a Thames Valley Group.centred on Stoke Poges. For further details contacthim at 'Pippins', Boulters Lane, Maidenhead, BerksS L6 8JT (Maidenhead 22445).

All of these groups are run on a purely voluntarybasis, therefore please remember that it is onlycommon courtesy to include a SAE when writing totheir organisers.

Southampton UniversityAn enthusiastic amateur computer club also exists atSouthampton University. They are building a 9900based system for Hardware Hackers (bolt holes willbe provided for all conceivable goodies) and are alsopurchasing a NASCOM kit so they can getsomething up and running quickly. As if all thiswasn't enough, they also publish an excellentmagazine; 'Benchmark'. Paul Maddison is thecontact, at 47 Westridge Rd., Portswood, Hants.

Finally, the ACC has now reached the end of itsfifth year, and the Annual General Meeting was heldon Thursday 13th April, 7.30 pm at the South BankPolytechnic, Borough Rd., near to the Elephant EtCastle.

For further details of the ACC, send a SAE to MikeLord, 7 Dordells, Basildon, Essex.

Developments in computing at Five WaysThe approach we are adopting to computing at KingEdward VI Five Ways School, Birmingham, that ofusing the computer as a problem -solving tool ratherthan to teach programming, is one that has arousedconsiderable interest among both representatives ofindustry and secondary education. We bring this toyour attention as we hope it may be of interest toyour readers.

The school currently possesses four terminalsaccessing the Newcastle-upon-Tyne Open Universitycomputer which we use extensively as an aid tocourse work in such '0' and 'A' level subjects asStatistics, Economics, Chemistry, Geography andMathematics. Some limited experiments in English,


French and Latin are also being undertaken. Inaddition to this work, every sixth -former, startingwith the present Lower Sixth, follows a half -yearcourse which aims to give practice in the approach toand solving a variety of problems, learning how thecomputer can assist him in this process, and it is thiscourse that has attracted the interest of localindustry. Representatives from BL Cars, Cadbury -Schweppes and Rubery Owen have visited theschool and, arising out of these visits, some jointventures are being planned. The first of these, a day-long business simulation exercise at the BL CarsSystems Training Centre at Studley, took place on 13March and again on 14 March.

The programs we use at Five Ways are either thoseon the Open University system or are written by ourprogrammers. These are boys who learn to programoutside school hours under the guidance of staff andsenior boys. Their training lasts for some two yearsbefore they start to write programs to the specifica-tion of subject staff, programs that, when tested,will be used in subject course work. The senioroperators are also responsible for the day-to-dayrunning of the terminal room.

If you would like to visit the school and take acloser look at what we do or to discuss our futureplans please do not hesitate to telephone.

For further information: Mr A F R Clements,Director of Computing, Mr H Reeves, Head ofGeneral Studies, King Edward VI Five Ways School,Scotland Lane, Bartley Green, BirminghamB32 4BT. Telephone: 021 475 8944 or 021 475 3535.*For 6800 users.Having grown rather tired of being the recipient of arather subtle advert every time my processor goesback into the BASIC command wait situation I

decided to amend the appropriate area of memory byusing a very small program written in BASIC. It is abuilding block for other tasks of a similar nature andcan be altered very easily from that shown below.

In line 10 the variable I is set to the first decimallocation at which the change is to be made. Line 20reads the first item in data which is the decimalnumber 84 which is in the ASCII code. A check ismade for a dummy number which, if equal to,branches to the end of the program. If not equal thenumber is POKE'd into the Ith location beingconverted into hexadecimal by the interpreter. Line50 then loops back for another number and so on.

It is not strictly necessary to have the dummy andcheck as the FOR/ NEXT loop would end before itgot there anyway but is useful to show it as underdifferent circumstances one could amend line 50 toread GOTO 20 and delete line 10. If you did this youcould almost amend line 20 to be INPUT A whichwould be the decimal equivalent of the hex characterand line 10 could be INPUT I. Line 45 would thenread LET I = I +1 in order to increment the locationnumber.5 REM THIS PROGRAM CHANGES 'MSI READY'

TO 'THE GHOST'10 FOR I - 349 to 35720 READ A30 IF A= 99 THEN 9040 POKE (I,A)50 NEXT I80 DATA 84, 72, 69, 32, 71, 72, 79, 83, 84, 9990 END

COLIN CHATFIELDMicro -Aid

Lloyds Bank ChambersCamborne, Cornwall

(SP) SIRTON PRODUCTS

We specialise in the S.100 Bus System with 8080 or Z.80 CPU's.

Board Kits8080 CPU Board with jump on reset complete

with sockets etc.4K RAM Board low power 450 n Sec. 121 L02-11Cassette Interface Board 187 to 540

Bytes/SecondPROM Board for 8K EPROM 2708 with Integral

ProgrammerS.100 Motherboard 10/11 slot with Active

Terminations

Coming SoonI/O Board Four 8 Bit Parallel and One Serial Port

- due mid -MayKeyboard AS R33 style having ASCII 7 Bit

output, with case - due mid -MayVideo Interface Board with 1K RAM

HardwareS.100 Edge connectors gold plated solder tailTransformer Pri 110/240v; sec. 8 v P 10 Amp &

30v. CT P 1 AmpBridge Rectifier with integral bracket rating

18 Amp

Integrated Circuits8080A CPU Chip 20 Sec21L02 RAM 1Kx1 Bit Low Power 450 n Sec21L02 RAM 1Kx1 Bit Low Power 250 n Sec2708 EPROM 1Kx8 Bit 450 n Sec1702 EPROM 256 x 8 Bit8212 I/O PORT 8 Bit

Few Only

£159.75

£125.75

E153.50

£139.95

£58.95

(98.25

£38.50T. B.A.

Please add 8% VAT + 30p post & packing (Transformer £1.00)

'Items thus marked also available ready built at extra cost.

NEW PRODUCTS BEING ADDED TO OUR RANGE EACH MONTH -WRITE OR 'PHONE FOR LATEST LIST AND PRICES.

SIRTON PRODUCTS,13 Warwick Road, Coulsdon, Surrey, CR3 2EF.Tel: 01-660 5617

£5.45

£12.25

£3.95

£11.25(1.50£1.95

£18.50£6.50£3.75

AIRAMCO LTD.MICRO COMPUTER PRODUCTS

741 op Amp MDIP 18555 Timer MDIP 30AY51013 UART 5:35AY52376 Keyboard Encoder 13:752102 1024 x 1 SRAM 1:35SPECIAL OFFER: 2102's 8 for 9:60

2104 4096 x 1 DRAM 16pin 3:99SPECIAL OFFER: 2104's 8 for 30:002107B 4096 x 1 DRAM 22pin 3:99SPECIAL OFFER: 2107B's 8 for 30:004116 16,384 x 1 DRAM 16pin 32:25

1702A 256 x 8 UVEPROM 5:902708A 1024 x 8 UVEPROM 12:252716A 2048 x 8 UVEPROM 29:908080A C.P.U. 11:50

S-100 BUS Mother Boards, Regulated Power Supply(10AMP or 20AMP), C.P.U. Power supply MemoryBoards (8K Static/8K, 16K Dynamic) PLUS manyothers.

Send LARGE S.A.E. for details

Discounts for quantity P.O.A.

Trade enquiries and orders welcome.P&P ADD 40p THEN ADD VAT @ 8%

BARCLAYCARD +

VISA MAIL ORDERONLYOwn We war /Mk

24Hr Ansaphone order service withACCESS or BARCLAYCARD No.

AIRAMCO LTD.30 WITCHES LINN

ARDROSSANAYRSHIRE, KA22 7BR

TEL. 0294-65530

65


EDITOR'S PlAYGROWID

THE SMALL SYSTEMAPPROACH TO

PROGRAM EDITInGPatrick Sutton

It is frequently stated that the approach to personal computing in Europe will be greatly influencedby the smaller amounts of money available for investment in computer hardware. To this end it islikely that the most popular machines in Europe will be small systems. It is my firm belief that thisfactor could favour development of systems which are far more user oriented and therefore better"personal computers".

The approach whereby the initial system is small The scroll editorand is only expanded when the need arises should, in In 1965 Mary Allen Wilkes developed what is knownmy opinion, be greatly encouraged. A small system as a scroll editor. (1) This program was originallydoes not necessarily mean a weak system. On the written on a Laboratory INstrument Computercontrary, I feel that by subscribing to large system (LIN C) which included a keyboard, displayapproaches, a fair proportion of the advantages oscilloscope, two randomly accessible tape drivesinherent to the personal microcomputer are lost or and only 2K of store. (2)obscured. A good example of this is found in the area Using the keyboard and 'scope interactively theof program development aids. I hope, in this article, programmer types in all corrections directly to theto illustrate this point by discussing a small system displayed manuscript. All editing commands, such aseditor. line feed and rub out, are delivered by single

This particular algorithm has been available since keystrokes. The major features of this scroll editor1966 but has been ignored by the majority of can be summarised as follows:mainframe manufacturers and, following in their 1) Characters can be inserted directly into thefootsteps, many microcomputer manufacturers. displayed manuscript from the keyboard. No

separate insert instruction is required.Line editors 2) At any position in the manuscript any amount ofThe vast majority of microcomputer based editors material may be added or deleted. The manuscriptavailable today (though not all, cf. the Research may be moved across the display in either directionMachines and PET editors) are line oriented. This and edited in any order.means that in order to edit the programme 3) The editor is unspecific and machine code up tomanuscript, numbered lines must be added or high level languages such as Basic may be editeddeleted. This is so even if the mistake occurs in only using the same algorithm.one character within a particular line. This type of 4) Line or sequence numbers are not stored in theeditor evolved due to the extensive use of TTY manuscript. A reference can be made to a lineterminals with mainframe machines. Here, single position using a program -maintained line counter,character editing is difficult to program because of and because of this approach any additions orthe obvious printing limitations. However, most deletions of lines are automatically accounted for.personal computer users today use visual display 5) User responsibility for manuscript maintenance,terminals (VDU's) and these printing limitations, ie, for example, packing of the characters, is removeddeletion of a character, are not a problem. Indeed, and is dealt with by the editor.those VDU's which are memory mapped, that is,those whose character positions on the screen areavailable as addresses in memory, are highly suitedto extensive text manipulations as I will mentionlater. Some of the character based editors that are The playground areaavailable require the programmer to manipulate the In order to understand just how this is accomplished,text with a series of commands, insert, replace etc. the reader should examine figures 1 to 4. These canWith some practice this approach can be very only convey the general principles involved, and theeffective. However, the approach discussed here is interested reader is directed to Wilkes's originalone which replaces editing commands by use of the . paper (Ref .1) in order to get more information in thevisually obvious. form of flowcharts etc.

66


Fig. 1. Creating the playground area.

Main manuscriptstring, on discor in memory

A

i 2 3 4

Editorialbuffer inmemory

N

POINT OFINTEREST

BLOCK MOVE FROM MAIN STRINGTO EDITORIAL BUFFER

1 A Al

3

POINT OFINTEREST

INSERTED PLAYGROUNDAREA

(Al is character initially adjacent to A)

Represents a functional block of characters i.e. 256 for a discsector or 40 for a VDU line length

The point of interest is located at point A in block 2. Blocks 1, 2, 3are moved to the editorial buffer in memory and a playground areais inserted between point A and the next character at point Al inblock 2.

Figure 1 shows that the editing procedure dependsupon the creation of a "playground" in themanuscript to be edited. The manuscript is kept ontape, disk, or in a portion of memory designated asthe main manuscript area. Upon selection of a pointof interest within the main manuscript, part of themain manuscript is transferred to an editorial buffer.Here an additional space, the playground area, isinserted between the point of interest and itsadjacent character. This additional space allowscharacters to be inserted into that portion of themanuscript contained within the editorial buffer, anddisplayed on the screen.

Insertion and deletion of characters are illustratedby figures 2A and 2B.

Fig. 2. Insertion and deletion of characters

2A. Insertion

A Al

3

tCHARACTER PLAYGROUND

INSERTED DECREASEDHERE IN SIZE

The character is inserted at point A. The playground is reduced insize by one character and point A moves one character closer topoint Al.

2B. Deletion.A Al

3

CHARACTERDELETED PLAYGROUND

HERE INCREASED IN SIZE

The character is deleted at point A. The playground is increased insize by one character and point A moves one character furtheraway from point Al. Note that insertion and deletion arecompensatory in their effects on playground size.

It is important to note here that in creating theplayground in the editorial buffer, adjacent blocks ofcharacters (in this case blocks 1,3) are alsotransferred to the buffer. This was included in theoriginal LINC program so that manuscript transfersfrom LINC tapes to memory would be infrequent, asthe tapes were slow, even if the point of interestmoved between various lines in the displayed

manuscript. If floppy discs are to be used, the sameapproach can be employed as disc transfers are alsorelatively slow.

But, if this editor is to be used in a situation wherethe main manuscript is held in memory, then theseadjacent blocks need not be loaded. In this casecharacter transfers can occur as quickly between themain manuscript and editorial buffer, as within thebuffer itself.

Fig. 3. Moving the point of interest

A. Forwards in the string

Playground moves forward

A Al

2

DIRECTION OF CHARACTERTRANSFER

To move forward in the string, characters are transferred frompoint Al to point A. The playground does not change size buteffectively moves forward.

B. Backwards in the string

Playground moves backwards

A AiiiDIRECTION OF CHARACTER

TRANSFER

To move backwards in the string, characters are transferred frompoint A to point Al. The playground remains the same size buteffectively moves backwards.

Figure 3 shows how moving the point of interest(which is shown on the screen by a cursor movingthrough the manuscript) is achieved. It can be seenthat by moving characters within the editorial buffer,the playground can effectively be inserted betweenany characters, so allowing extensive editing.

Should the next position of interest in themanuscript lie outside the editorial buffer, then thebuffer must be reloaded. Figure 4 shows how this isaccomplished moving in either direction and also,indirectly, the true power of the playgroundapproach.

Fig. 4. Moving to characters not contained in the editorial buffer- block transfers

4A. Forwards

BLOCK MOVE

EDITORIALBUFFER

Main manuscript string

A Al

4

BLOCK MOVE

CHARACTERTRANSFER

To move forwards, block 1 is transferred back to the main stringand block 4 is transferred to the editorial buffer. Charactertransfers occur as in Fig. 3A to move the playground to the pointof interest.

NoteThe original boundaries of block 1 (i.e. which characters are in

block 1) could change during editing. The important point to noteis that the same number of characters are transferred back to themain manuscript, as are being moved to the editorial buffer fromthe manuscript. Any characters overwritten in the main string, willbe contained in the editorial buffer. Thus the playground iscontinually concerned with containing characters which havebeen added to the string thereby increasing the string's overalllength.

67


4B. Backwards

BLOCMOVE,

Main manuscript string

4

A Al

2 3

CHARACTERTRANSFER

BLOCK MOVE

}EDITORIALBUFFER

To move backwards, block 3 is transferred back to the main stringand block 0 is transferred to the editorial buffer. Charactertransfers occur as in Fig. 3B to move the playground to the pointof interest.

It has been stated previously that line numbers donot form part of the manuscript. The manuscriptconsists of a character string which contains nocontrol information other than line terminators.Thus, the division of the manuscript into blocksdefines only the number of characters, and not theirfunctional relationship. When the transfer of block 1back to the main manuscript occurs (fig. 4a),different characters - ie edited manuscript - maybe transferred back. On examining the figures care-fully we can see that any information which is trans-ferred to the main manuscript will over -write onlyinformation which is already contained in theeditorial buffer. This is so because of the way theplayground area is created, and maintained. Allinformation that is required - that is, not edited outis retained during these transfers.

Insertions and deletions are in generalcompensatory in their effects on playground size.However, should there be more insertions thandeletions the total manuscript must increase inlength.

In fig.4a, as block 1 is transferred to the mainmanuscript string, block 4 is transferred to thebuffer. Any increase in the text length due to"excess" insertions will be accounted for by adecreased playground area. On completion of theediting process the original manuscript will berewritten by passing it through the editorial bufferand therefore across the playground area (fig. 4a).When all the text has been passed through the bufferand back to the main manuscript, the playgroundwill effectively be pushed off the end of themanuscript. The important point to note here is thatthis total rewriting is only done at the end of editing,and not after each insertion.

The exception. One exception to this rule ofrewriting only at the end of editing is when the play-ground area fills up. If this happens, the main manu-script is rewritten via the editorial buffer until theplayground area is pushed off as before. When this iscompleted, the point of interest is relocated using theline counter, and a new playground is created as infig. 1.

Wilkes has commented (1) that the minimum sizeof the playground area for efficient editing should beabout five times the length of a typical manuscriptentry. For instance, when programming in assembler

68

language a typical entry (line length) will be about 15characters, and so the playground area should be atleast 75 characters long. Programming in Basicshould require a larger playground, whereas to edit atypical machine code entry would require a smallerplayground. Obviously, by choosing the larger play-ground all texts from machine code to Basic could beedited efficiently by the same program.

Small system powerIn my opening comments I said that I would use theexample of this editor algorithm to show that smallsystems can in fact be powerful. In the light of thiseditor approach let us now look at two currentlyavailable systems: the NASCOM 1 and a newcomer,the MICROS.

Both these systems use the Z-80 microprocessor,and both have a keyboard with a memory mappedVDU included in the basic kit. This configuration,built originally as an economical approach topersonal computing, is ideal for this powerful andversatile editor. The Z-80 has a single instructionblock move, useful in this case for buffermaintenance. The memory mapped VDU provides anexceptional editorial buffer and playground facility.Combining these two factors would alone providethe user with single character editing, of anadvanced nature, with its associated ease ofprogramme development - in the minimum systemoffered!

In order to achieve this with a minimum system,both the editor and the main manuscript wouldreside in the random access memory (RAM), whilstthe video RAM would be the editorial buffer andplayground.

As mentioned earlier, if memory is used to containthe manuscript then adjacent blocks of charactersneed not be loaded into the editorial buffer, thusreducing the amount of store required. In the case ofvideo RAM, adjacent blocks of characters would bethose required to display lines of manuscript, bothbefore and after the line containing the point ofinterest.

In applications where large amounts of text are tobe edited, the system may be expanded by addingmore memory or by adding a disk. If a disk is added,then by storing the main manuscript on it a virtuallyunlimited length of text may be edited using only 2Kof RAM. It is important to stress here that, in thecontext of my opening remarks, a disk is notessential to start editing. I believe that given the Z -80's powerful instruction set and the memorymapping of the video display, a good text editor plusa fair sized manuscript would fit into the 2K ofmemory available in these basic kits.

I hope that this article has gone some way towardshowing that small well developed systems can bevery powerful. I also hope that users of micro-computers in Europe will not fall for the inefficienciesfound in many American systems, where in order todevelop even a short program at least 8K of store isneeded for the development software. It is smallwonder that, with this approach to software,megabyte systems are frequently encountered.

References1. M.A. Wilkes, Scroll Editing: An On-line AlgorithmFor Manipulating Long Character Strings; IEEETransactions On Computers, Vol. C-19, No 11, Nov1970, pp1009-15.

2. W.A. Clark and C.E. Molnar, A DescriptionThe LINC; in, Computers in Biomedical Research.Vol 2, R.W. Stacey and B. Waxman, eds., NewYork, Academic Press, 1965, pp35-66.


ETHE EUROPA BUS

A MICROPROCESSOR BUSSPECIFICATION FOR

EUROCARD

What is a bus?In its simplest form a computer may be considered toconsist of three parts, the central processor (CPU)which does all the hard work, the memory, which isin effect a large number of individually labelled'holes' in which data or instructions can be stored,and the input and output devices (I/0), throughwhich it can talk to the outside world. For the systemto work the CPU must be able to communicate withthe I/O devices and the memory, and in some casesit will be necessary for the I/O devices to transmitdata directly to and from the memory. Potentially,individual connections could be made between thevarious devices and the CPU, but this would requirea lot of wires, buffers, edge connector pins and soon. A much more satisfactory solution is provided bythe concept of the bus, which has become anessential part of microprocessor technology, bothinside the devices and in functional systems. Inessence a bus is a collection of wires which areconnected to and shared between all of thesubsections of the system. For a computer systemthe bus may be considered in three parts.

Firstly there is the address bus, which is used bythe CPU to tell the rest of the system which item ofdata or which I/O device it is interested in. In simplesystems this is a uni-directional bus, that is to say it isalways "driven" by the CPU, other componentssimply "listening for their name to be called".

The second bus is the data bus. This is used totransmit data around the system, and is therefore ab/ -directional bus. It is driven by the CPU when it issending data to the memory ('writing'), while it willbe driven by a memory or I/O card when the CPU isasking for data ('reading').

The final ingredient necessary to get all this towork properly is a set of signals, known as thecontrol bus, to make sure that everyone knowswho's doing what and when. As an example, thefour most important control signals in the busdefined here are MREQ, IORQ, RD and WR, all ofwhich will normally be driven by the CPU. MREQtells the system that the CPU wants to talk to thememory location corresponding to the address onthe address bus. IORQ similarly shows interest in anI/O device (obviously MREQ and IORQ should notboth be active at the same time. RD tells thememory or I/O device that the CPU wants it to put its

Bob CottisPippins, Boulters Lane,Maidenhead, Berks.

Dave Howland186, Courthouse Road,Maidenhead, Berks.

Pat Crowe22, Ringsbury Close,Purton, nr. Swindon,Wilts.

current data onto the data bus. Similarly WR tells theaddressed memory or I/O device that the CPU hasput new data for it onto the data bus.

Why do we need another standard?There are many ways in which a microprocessor buscan be controlled. For example, 6800 based systemsoften combine the functions of RD and WR definedabove into a single signal. This tells memory whetherto read or write when MREQ goes active. Theobjective of any bus specification is to define amechanical and electrical system such that printedcircuit cards can be produced to run in any systemwithout problems. The two buses which haveachieved widespread use are the S-100, specified in1974 by MITS for their Altair computers, and theS-50, defined by South-West Technical ProductsCorporation for their 6800 based machines. A largenumber of standard parts are available for thesebuses, so why do we need another standard?

The main disadvantage of both these buses is thatthey were designed around one particular processor,and all the quirks of these machines have been builtin. It is possible to squeeze other processors onto theS-100 and S-50 buses, but this inevitably involvescompromises. Anyone who has seen the "trailers"gradually being released for the new breed of 16 -bitmicroprocessors due out this year will be aware thatthe amateur is very soon going to have access tomicros with the power of today's minis. These

69


devices can probably be accommodated on the 8 -bitS-100 and S-50 buses, by transferring two 8 -bit bytesfor each 16 bit word, but this will s/ow the processordown by a factor of two. Additionally, without extrahardware the standard buses will only accommodate64K bytes of memory, whereas the Intel 8086 isreputed to be able to address 1M bytes, and the ZilogZ8000 8M bytes! At present the cost of filling 64Kbytes may seem sufficient to make this problemrather irrelevant, but a price of £4 per chip does notseem unreasonable for 16K dynamic RAMS in a fewyears' time, which will change things somewhat.

Thus the primary aim of this specification is todefine a system based on readily availablecomponents, which will have sufficient generality toaccept all current processors, and hopefully also thedevices which wil I be available in the near future (in15 years we probably won't need a bus, just take a 40pin chip, connect to + 5V, a keyboard and a TV andyou'll have the power of an IBM 370 at yourfingertips).

ObjectivesWhen the outline of this definition was firstproposed, it was suggested that it should: -

1) Use readily available equipment practice(boards, card -frames, connectors etc., andallow easy manufacture of printed circuitboards by the individual if required.

2) Be sufficiently flexible in definition to allow itsuse with any existing (and hopefully future) 8bit microprocessor without penalizing any oneby requiring complex logic to interface with thebus.

3) Be sufficiently flexible to allow goodpossibilities for implementing personalexperimental or non-standard systems withoutdeparting from the framework of the standard.

4) Be flexible enough to make it possible toproduce equipment to full professionalstandards while at the same time not penalizingthe experimenter with a low budget.

Subsequently it became clear that many more 16 -bit processors would soon be available at reasonableprices, and the bus definition was modified to allowfor 16 data lines. This may at first seem something ofa waste of 8 valuable pins for people using 8 bitmachines, but it should make it possible to switch toa 16 bit machine with very little difficulty. Devicessuch as memory mapped VDUs and cassetteinterfaces can be used directly by 16 bit machines,simply by writing the software to ignore the top 8bits. Memory cards can be used in pairs by assigningone to the bottom byte, and the other to the top.

BoardsThe board size selected is that known as DoubleEurocard. The dimensions of this board are shown infig. 1. The Eurocard is rapidly becoming theEuropean standard, and even the U.S. is beginningto succumb to its Teutonic charm. Consequentlyequipment such as racks, cases, prototyping cardsand connectors are readily available from manysources. The smaller Single Eurocard can be usedwith the same connectors and bus convention, but itis not recommended, as considerably less than halfas much useful logic can be fitted on the board, forthe same expenditure for connectors and busbuffering.

ConnectorEurocards can be used with standard edgeconnectors or with two-part 'indirect' connectors.The two-part, 64 -way connectors have beenspecified for the following reasons,

1) They are highly reliable and, in the unlikelyevent of damage, replaceable.

2) They make it possible for the amateur toproduce reliable printed circuit boards withoutneeding access to gold-plating equipment.

3) They are not all that cheap (about £5 per pair),but the circuit boards will cost a pound or twoless because they do not require to be goldplated.

4) They appear to be the "industry standard"connector for Eurocards, and are readilyavailable.

One 64 -way connector provides just enough pinsfor the bus. The other connector is optional, and maybe used for input and output connections, or for anyother functions the user wishes. Memory cards willnormally be fitted with only one connector.

Two versions of the 64 -way connector areavailable, referred to as 64/64 and 64/96. The firstnumber refers to the number of pins actually fitted tothe connector, while the second number is thenumber of pin positions in the connector housing.The most significant difference between the twoversions is the spacing between the two rows of 32pins each. This is 0.1 inch for the 64/64, and 0.2 inchfor the 64/96. This makes it much easier to produce aprinted circuit backplane for the 64/96 connectors,and for this reason they are preferred. Howeverprinted circuit boards can and should be designed toaccept either connector (see fig. 2).

Viewing the board from the component side, withthe connectors on the right, the bottom connector isthe bus connector. The connectors are mounted onthe component side of the board.

ElectricalThe pin assignments for the various bus signals aregiven in Table 1.

Five basic signal types are used,

70


1) Power Supplies Two pins are used for eachpower supply (four for ground). This permitsprinted circuit backplanes to be designed withfairly solid power conductor traces. Theconnector rating is 2A per pin, so the maximumcurrent which should be drawn by any one supplyof a card is 4A. Ground and +5V are situated atthe ends of the connector so that the traces forthese supplies can be made a lot wider.2) Tri-state Lines Used for synchronous signals

which may be driven from several sources. Maybe driven by LS, standard or Shottky TTLbuffers. LS buffers are recommended for mostsystems. Each card should preferably presentonly one LS load to the bus.

233.4mm

160 rm

Component

Side

Fig. 1 Double Eurocard

Optional

Connector

Bus

Connector

3) Daisy Chains See below for more details onhow these work. On cards not using thesesignals the in and out pins should be connectedtogether. On other cards the output should bedriven by one standard TTL output or theequivalent. Inputs should be TTL compatible.

4) Lines Driven From One Source Only Usedfor the system clock(s). Driven by permanentlyenabled tri-state buffers or the equivalent.

5) Open -collector Lines Used for asynchronoussignals which can be driven from more thanone source. Should be driven by devicescapable of sinking at least 16mA.

Signal PolarityAs noted above all signals should be negative true.This means that a 'I' on the address or data buses, oran active control signal will be signified by a TTL lowvoltage. This convention is indicated by the bars overthe signal names. This convention should be strictlyadhered to for control signals as it ensures that theywill be pulled up to an inactive state when the controllines are not being actively driven.

For the address and data signals the use of anegative true convention means that cheap inverterssuch as 74LSO4 or 74LS14 will give positive truesignals on PROM and I/O cards, which will make lifesomewhat easier. In general however the individualwill experience very few problems if a positive trueconvention is preferred for the address and datasignals.

The Bus Signals1) AO to A15 These constitute the address bus.

Al5 is the most significant bit.2) DO to D15 These constitute the data bus. D15

is the most significant bit. 16 bits have beenspecified to allow for existing and future 16 bitmachines. 8 bit systems can use D8 to D15 forother functions, but bear in mind that this maymake it difficult to change to a 16 bit processor.

3) BAI, BAO, 1E1, 1E0 These signals are to beused for daisy -chained signals (i.e. 1E0 fromone card goes to IEI on the next). Theseparticular signals are for the Z80, being used tocontrol bus access priority during DMA, and tocontrol interrupt priority. Thus, when IEI to a

ROW c

ROW b

ROW a

4)

5)

Fig. 2 Pad Configuration to Accept 64/64and 64/96 Connectors

card is low any interrupt logic on the card willknow that no higher priority interrupts arepending, and it will generate an interrupt ifrequired. At the same time 1E0 will go high totell lower priority cards that its interrupt ispending. Any system requiring daisy -chainedlines should use these pins. These signals willnot normally be used by memory cards, andthese four pins may be used to extend theaddressing range to 1Mbyte by using a 'split'

bus, with memory in one section andperipherals in the other.Clocks These signals allow for two systemclocks. For machines which generate a signalto indicate that the first byte of an instruction isbeing fetched, this should be placed on pin15a.Otherwise this pin can be used for the lessimportant system clock in a two-phase system,or for any other function. Pin 15b/c should beused for the most important system clock (i.e.02 for the 6800 and 8080).DWAIT This signal is used to hold theprocessor via the WAIT line, and disable theCPU card tri-state buffers. This signal may beused by a single card dynamic RAM controllerto provide for dynamic RAM refresh, ofteninvisibly. It is also necessary for theimplementation of virtual memory (there is notspace here to describe this technique, seeBYTE Vol. 2 No. 7 for an introduction).

71


Pin

Table 1. Bus Signals12)

Row a Row b/cSignal Type Note Signal Type Note

1 Ground 1 Ground 1 13)2 Ground 1 Ground 1

3 +12V 1 +12V 1

4

5

-12V 1

77 2

-12V 1

214)

6 Al 2 2

7 AT 2 AS 2

8 A6 2 R7 2

9 TB 2 2

10

11

A10 2

AT7 2

All 2

rr7 2

12 ATT 2 Al! 2

13 BAI (TIT)3,2 3 BAO (A1 /) 3,2 3

14 /EI )3,2 3 TEO (T77) 3,2 3

1516

/x1M1 (P) 2,4 4

5

4 4

6 15)DWAIT 5 INTAK 2

17 DMAR 5 7 DMAG 4 7

18 8 INT 5 8

19 MREQ 2 9 WR 2 11

20 7771T 5 13 WAIT 5 14

21 IORQ 2 10 hb 2 12

22 INHR 5 15 RFSH 2 16

23 E 2 2 M 2

24 B7 2 2

2526

FT 2

FT 2

177 2

157 2 16)27 BB 2 DZS 2

28 DIO 2 Dll 2

29 BT7 2 5T 2

30 TIT 2 D15 2

31 -5V 1 -5V 1

32 +5V I +5V

6) INTAK This signal is generated by the CPU toindicate that it has accepted an interrupt.

7 DMAR and DMAG These signals are used tdcontrol bus access by peripheral devices usingdirect memory access (DMA). DMAR isgenerated by the peripheral to indicate that itwants to use the bus, while DMAG is used bythe CPU to signify that the request has beengranted.

8) NMI and INT Non-maskable and standardinterrupt request inputs. The method ofresponse to an interrupt will inevitably beprocessor dependent, but if possible it issuggested that the CPU should be configurableto respond to an interrupt consisting solely ofINT going low until INTAK goes low. This willpermit peripheral cards requiring single levelinterrupts to be produced in a non -processordependent form.

9) MREQ This signal signifies that a memoryaccess is in progress and that the address buscarries a valid address. For the Z80 MREQ canbe used directly, while for the 6800 thesuggested logic is (VMA. 02.10BLOCK), whereIOBLOCK indicates that a memory mappedinput-output block of 256 bytes is beingaccessed.

10) IORQ This signifies that an I/O access is beingperformed, and that the lower byte of theaddress bus carries a valid I/O address. For theZ80 IORQ can be used directly, while thesuggested logic for the 6800 is VMA. 02.IOBLOCK.

11) WR Write strobe, indicating that the data buscarries valid data to be written to memory orI/O. For the Z80 WR may be used directly,while the suggested logic for the 6800 is VMA.02.R /W.

RD Read strobe, indicating that the addressedmemory or peripheral should place data on thedata bus. For the Z80 RD may be used directly,while for the 6800 VMA. 02.R/W may be used.

RESET Resets processor and peripherals.

WAIT Used by slow memories and peripheralsto delay processor read or write cycles. TheWAIT line should not normally be held low forextended periods as this could cause refreshproblems for dynamic memories andprocessors such as the 6800.

INHR Used to inhibit reads from RAM boardsto allow 'shadow' PROMS to be temporarilyplaced at the same memory locations duringsystem initialization. Does not prevent writingto RAM.

RFSH This signifies that the lower orderaddress bits carry a valid refresh address for useby dynamic memories.RFSH.MREQ may beused for the Z80, while a separate controllerwill be required for the 6800.

Control logic for specific processorsIt can be seen that the above specification usescontrol signals which are similar to those for the Z80.There are several reasons for this, the major onebeing that of adaptability. Memory boards willnormally use MREQ to gate the selection of memorychips, RD to enable the data bus buffers onto thesystem data bus, and WR to write to the selectedchip(s). Using this convention the actual timings forMREQ, IORQ, RD and WR will not be critical to thelogic of the memory card, providing the address isvalid throughout MREQ pt IORQ, the data is validthroughout WR, and the CPU does not try to drivethe data bus during RD. To give a safety margin,address and data buses should be valid for at least5Ons before and after the relevant control signals areactive. Other schemes which effectively use only twosignals to access memory (e.g. VMA and R /W forthe 6800 or MRD and MWR for the 8080) causeconsiderably more problems, particularly forprocessors, such as the Z80, which generate WR latein the memory access cycle.

Despite this similarity to the Z80, it must beemphasized that this is not just intended to be a Z80bus. The detailed logic has been worked out for the6800 and the Z80, and it has proved possible to makethem sufficiently compatible that static memorycards and simple I/O can be directly interchangeable.The detailed logic for other processors has not yetbeen defined, as the authors of this standard do nothave the first-hand knowledge necessary to definethe control logic in the optimum way. We havehowever checked that other processors can, as far aswe are aware, be used with this bus structurewithout serious problems. The second connector canbe used for additional signals such as the CRUfunctions for the 9900 (if they can't be fitted into thebus itself).

It is hoped that user's groups for these otherprocessors will draw up their own definitions forsuitable logic. The authors of this standard would bepleased to give any assistance possible in theformulation of thses definitions.

We would also be pleased to receive comments,criticism and suggestions for improvement, eitherdirectly, or through the pages of PCW.

72


Products PageKIM -1, America's most popular low costmicrocomputer system card, is now available fromPELCO (ELECTRONICS) LTD., Enterprise House,83/85 Western Road, Hove, Sussex (telephone:(0273) 722155), at £185,00, plus V.A.T.

The product is ideal for the enthusiast who wishesto get down to the real business of microprocessordesign, which is the software. The Card comes builtand tested, and only requires the addition of a 5 voltpower supply, and the user can immediately startprogramming. Based on the third generation 6500microprocessor, which itself was developed from thepopular 6800 product line, all the powerful softwareand hardware features associated with this processorare available to the user. In addition, the Card comescomplete with interface circuiting to enable a simpleconnection to a low cost audio cassette player,which can then be used to store user programmes.

The Card also contains 15 additional I/O lines forconnection to external circuiting, as well as apurpose built teletype interface for engineers withsuch facilities.

The basic Card uses a ROM -resident editorprogramme with de -bug facilities such as single-stepability built in. With 1K bytes of RAM easilyexpandable to 4K bytes, and with the ability toexpand to a full 65K bytes if required, the uses of theproduct are virtually limitless as far as currentmicroprocessors go. Support for the product isavailable both from the various 'user' clubs andpublications, and from Pelco (Electronics) Ltd., whoare Rockwell's Distributors and Representatives in

facilities to all users of the wide variety of Rockwellmicroprocessors from their laboratories located inHove.

THE MICRONICS COMPANY have released theBritish designed and assembled MICROSmicrocomputer system. The price of £550 includes aZ 80 cpu, debug and monitor ROM, 2K bytes RAM,audio cassette interface and checksum errordetector, video monitor interface, UHF TVmodulator, 53 key contactless ASCII keyboard,power supplies and cabinet. RS232 and 20 MAcurrent loop outputs are provided for a teletype or amatrix printer and the RAM is expandable up to 64Kbytes. Options under development include a

minifloppy disc, BASIC compiler in ROM andcassette, NPL time decoder, TELETEXT decoder anda selection of cassette games.

The system has been designed to meet the needsof educational establishments, personal computing

and small business users at a realistic cost and is alsoavailable in kit form. A data sheet is available onrequest and enquiries from common market andoverseas agencies would be welcome.

THE MICRONICS COMPANY1, Station Road, Twickenham, Middlesex.

The introduction of the Equinox 100 computersystem at the Mini -Micro Exhibition, held recently atthe U.S. Government Trade Center in London makesavailable, for under £3,000, a system based on thepopular 8080 or Z80 microprocessors with 32K bytesof read/write memory and dual floppy disc drives.

The Equinox system is based on the industrystandard S-100 bus, which supports the largestvariety of plug -compatible boards and peripheralsavailable. More S-100 bus based computers arecurrently shipped than any other system, with anestimate of over 50 manufacturers in the U.S.making compatible boards. These include CPUboards based on the Intel 8080, Intel 8085, Zilog Z80,and the 16 -bit Western Digital MPU, as well as awide range of memory boards. In addition, there aredigital to analogue converters, speech synthesisers,speech recognisers and a variety of other interfaceboards, all making computing power generallyavailable at minimal cost.

The special features of the Equinox include aconstant voltage power supply, to overcomefluctuations in line voltage, and the Wunderbussmotherboard, designed by the famous Americanelectronics engineer, George Morrow. This busboardhas both full shielding and active termination to keepsignals "textbook clean". The size of the mainframepermits expansion to a low cost multi-user, multi-tasking, time sharing system.Further information is available from Interam ComputerSystems Ltd., 59 Moreton Street, Victoria, London SW1. 101-834 0261/2733)

AMI-COS Microprocessor SystemIn conjunction with INCAA Computer Design BureauRITRO offers the modular AMICOS microprocessorsystem. Other currently available systems do nothave, in general, universal extension possibilities,whereas the AMICOS system can be configured tothe users' exact requirements, thanks to modulardesign and an effective bus structure.

AMICOS is based on the well known S6800 familyof American Microsystems Inc. and is offered in kit -form, with PC boards on Eurocard format. Eachmodule comes with extensive documentation onassembly and applications; and with the first moduleRITRO supplies an attractive binder with data on theS6800 and general information on AMICOS.

Modules Available or Planned for the Future

* THE CENTRAL PROCESSING UNIT, with a fullybuffered S6800 MPU, clock circuitry and memory.

* A CONSOLE DRIVER with operating system inROM. This module includes switches,pushbuttons and LED's, forming a binary console.With the above two modules the user may beginto get acquainted with microprocessors and study

73


the instruction set by running small programs viathe console.

* THE CPU CONTROL CARD. Through this modulethe CPU selects the desired I/O card and handlesthe various interrupts.

* 8 -bit parallel I/O module with PIA and decoding.* Problem oriented modules such as model train

controlling, morse generation and decoder, A to Dconverter, TV -driver, audio -cassette driver, etc.,

* Memory extension (RAM Et E -PROM).* Various output cards (opto-couplers, relays).Details from RITRO Electronics (UK) Ltd., Grenfell Place,Maidenhead, Berkshire.

This portable video terminal produced by ProntoElectronic Systems, offers a cost effective alternativeto the standard teletype, by using the latestmicroprocessor technology. It gives the computersystem or microprocessor user a no -fuss method ofdata entry at a fraction of the cost normallyencountered.

The Model 601A is built into an attractiveexecutive -styled carrying case, measuring only16}" x 12i" x 5". It has been designed to interfacewith a video monitor or modified home TV, and themicroprocessor or computer system being used via a20mA loop (or RS232 option).

Applications for the Model 601A are extremelyvaried, but one important area is as a low costinterface to microprocessor and computer systems,and for educational and commercial applications. Italso offers the opportunity for establishing remoteterminal or message display facilities at suchlocations as conferences, exhibitions, businessmeetings or even in the home.

Further details: Pronto Electronic Systems Ltd. 645-647 HighRd, Seven Kings, Essex IG3 8RA.

Tap -a -dipNew from the American manufacturers of EZ Hooksis a test clip to simplify testing on dual inlinepackages - the TAP -A -DIP.

TAP -A -DIP is presented in 14 -line and 16 -linemodels. Critical width does not change duringinsertion and removal, making TAP -A -DIP ideal forhigh density boards. End contacts lock firmly undereach of the end leads, and by this strength TAP -A -DIP is able to double as an insertion and removaltool. One piece square contacts extend above thetop of TAP -A -DIP to enable EZ Hooks to be easilyconnected.

As with all EZ Hooks, TAP -A -DIP is represented inthe UK by Special Products Division of BritishCentral Electrical Co. Ltd., New Street, Ringwood,

Hants., from whom further information may beobtained.

A step-by-step approach to mastering micro-processors, by ERAERA has introduced the MP Experimenter, a trainingcourse utilising a simple but highly effective step-by-step procedure which commences with anexplanation of basic logic elements and graduallybuilds up to an appreciation of a full microcomputer.

Central to the course is a portable desk -topmicrocomputer unit incorporating a control panelwhich is used in conjunction with a set of overlaysdesigned to guide the student through each teachingstep. The system includes an operating manual and afive-volume set of course notes containing a wideselection of practical experiments. This enables theuser to teach himself at his own pace in the home,office or laboratory.

Aimed primarily at technologists and engineers,the system will also be invaluable to managers andexecutives who have perhaps been overtaken by thisfast moving technology and can now bringthemselves up-to-date using a method which fits inwith their busy day-to-day schedules.

The MP Experimenter hardware is programmed tosimulate each part of a microprocessor in turn,providing the following functions: -

* An 8 bit adder/subtractor* An 8 bit arithmetic logic unit (ALU)* An accumulator* An accumulator with data memory* A simple microcomputer* A Hypothetical microcomputer* An 8080 microprocessor system with a monitor

Further information from:Chris Nabavi, Computers and Automation , ERA Limited,Cleeve Road, Leatherhead, Surrey, KT22 7SA. Telephone:Leatherhead (037231 74151 or Hugh Cullimore Worplesdon(048631) 2729.

74


Gamma Computer Products, a new membercompany of the Nottingham -based GammaAssociates Group, has introduced the GammaProteus Compact (GP - Compact) microcomputersystem .

The powerful desk top system which is making itsworldwide debut in the UK market is completelyportable. It incorporates a 64K byte DEC LS1/11cpu, a Burroughs "Self Scan" Plasma Display, aHoneywell "Microswitch" keyboard, a Pertec'Microfloppy" floppy Disk and a Centronics Series

700 printer. The total product comes from the GeneralRobotics Corporation of Hartford, Wisconsin.

Price of the 64K byte system - the maximummemory that the DEC LS1/11 will 'support - plusthree floppy drives is £7,500 - one off - with a fullsoftware licence. Susbstantial quantity and OEMdiscounts are available which can reduce the price toalmost the £5,000 level.

The complete sytem is extremely compact in size,being 41 cms high, 53 cms wide and 61 cms in depth.It weighs under 22 kgs.

The new system offers users all the softwarefacilities of the DEC LS1/11 system, includingBASIC, a Fortran IV compiler and APL. It also offersTSX - a timesharing executive which allows anumber of users to timeshare the background jobfacilities of the RT-11 operating system. The systemwill also run RPL - the PDP-11 version of theNational Computing Centre Filetab language.

The company also expects to sell to OEMs whowill be able to tailor systems for small serviceoperations, such as accountants, solicitors,architects, consultants and small distributors andmanufacturing companies.

DECstation 78 from Digital Equipment Co. Ltd.,Digital House, Kings Rd. Reading, Berks.DECstation78 is the new LSI version of Digital Equipment'sfamous 12 -bit PDP-8 family of computers.

The compact, inexpensive DEC station 78 hasbeen designed as an effective, problem -solving toolfor a broad range of user environments - everythingfrom personal computer and fixed functionapplications to real-time, multitasking operations andnetworking.

The PCM-12 is an inexpensive small computer that isfully software -compatible with the Digital EquipmentCorporation PDP-8 family of minicomputers.

The PCM-12 is designed around the Intersil IM6100microprocessor, a device which is software -compatible with the most popular minicomputer everdeveloped - DEC's PDP-8. That means that thePCM-12 can execute nearly all PDP-8 software,without modification. And that includes assemblers,editors, debug routines, interactive languages likeBASIC and FOCAL and scientific languages likeFORTRAN. With the addition of a floppy disk orother mass -storage device, the PCM-12 is evencompatible with DEC's sophisticated OS -8 operatingsystem.

Largely because of its PDP-8 compatibility, thePCM-12 has already been widely accepted into OEM,educational, scientific and home computingapplications. PCM-12's are being used today inbusiness data processing; machine tool control;national defense applications; energy development;physiological, geological, chemical, vibration, laser,and medical experimentation; dedicated IM6100microprocessor system development; and homecomputer games.

The IM6100 microprocessor is essentially an LSIversion of the CPU in the P DP -8/ E, the latest versionof the classic PDP-8. This chip makes it possble forthe PCM-12 to perform like a PDP-8/ E in nearly everyrespect - and even provide features that the '8lacks. Features like a separate control panel memorywhich makes the entire main memory available forthe user's program. And a built-in binary bootstraploader - you simply flick a front panel switch to loada machine language program from paper or magnetictape. The PCM-12 control panel provides essentiallyall P DP -8/ E functions, plus single -clock anddecrement -address functions.

With provision in its design for vectored priorityinterrupts and direct memory access, the PCM-12 isa true minicomputer designed around a powerfulmicroprocessor. The machine's bus -orientedarchitecture provides 15 identically operating cardslots - enough room for up to 32K words of memoryand a full complement of peripheral interfaces.

The PCM-12 is a solidly -designed general purposecomputer that can serve as the heart of the simplest orthe most sophisticated small computing system.

Electronic secretary of the yearThe new Mind Reader personal computer is fromOptimisation, 45 South St. Bishop's Stortford,Herts. It plans, files and reminds. No more missedappointments, forgotten birthdays, double luncheondates and late meetings. Its computer memory solvesall these problems and further uses its advancedelectronic technology to display the time and date.It's also a stop -watch, electronic calendar and two-time zone clock and calendar.

The Mind Reader will manage appointmentsschedules and maintains four categories in which tofile special messages, displaying them at theappropriate date and time.

It will also display a daily schedule - today,tomorrow or any other day. At the touch of a key'things to do' are on file for each day. The MindReader will retail at around £300.

75

FEATURING

Systems for smallbusinesses

Computers in Education

Computers in the home

Amateur computing

ATTRACTIONS

Competitions

Best home brew system

Best software

Best school applications

Best home applications

PLUS

Famous Personalities

Innovators corner

Video games

PCW Micro chess championship

Seminars, PCW consultants in attendance

compelecelectronicsWelcome you

to theAltairAge

THE MICROSYSTEM FOR THE PROFESSIONAL(System 9V illustrated - £6,781 to £7,026)

We supply a complete range of low-cost Altair systems andsoftware for educational, industrial and business use.

Main FramesChoice of front panel or TurnkeyMemory capacity to 64K bytes

Bulk Storage Devices10 megabyte disc system; standard floppy disc system; mini floppydisc system or cassette.

InterfacesA variety of plug-in analogue, digital and industrial control inter-faces supplied from stock.

TerminalsSoroc VDUs, Centronics and Qume line printers

SoftwareFORTRAN, BASIC, multi-user BASIC, disc operating systems,assemblers, editors, monitors, business packages, word processingpackages, etc.We supply fully commissioned Altair computer systems complete withterminals and software.

Telephone 01-328 1124 for a Demonstration

compelec electronics limitedCompelec Electronics Limited, 107 Kilburn Square, London NW6. Tel: 01-328 1124

uuC 1111ITTEN

PA% IPBRITAIN'S LONGEST ESTABLISHEDCOMPUTER SHOPHARDWARE

COMPUTER MP -68 with 4K Memory Et Control Interface Kit £275.00 Assembled£330.00MEMORY MP -4K Memory Kit £70.00 Assembled £90.00 MP -8 8K Memory Kit£170.00 Assembled £195.00DISC SYSTEM DMA1 1.2 megabyte twin system Kit £1450.00 Assembled£1525.00MF-68 160K byte twin minifloppy Kit £800.00 Assembled £860.00TERMINAL ASR43 Upper/Lower case 132 Char. 30 cps KSR Assembled £915.00CT -64 Upper/Lower case visual display terminal Kit £230.00 Assembled £315.00CT-VM Video Monitor for CT -64 Assembled £140.00PRINTER RICOH RP -40 Daisy wheel printer 30 cps Assembled £1800.00CENTRONICS 701 Dot Matrix printer 60 cps Assembled £1400.00PR -40 40 column tally roll printer 55 cpm Kit £200.00 Assembled £250.00CASSETTE AC -30 Cassette Interface for 2 audio cassettes Kit £65.00Assembled £100.00GRAPHICS GT-61 Graphics Terminal with P197 Power SupplyKit £92.50 Assembled £120.00SUNDRIES MP -R EPROM Programmer for 2716'sKit £36.00 Assembled £43.00MP -N Calculator Interface Kit £37.00 Assembled £44.00MP -T Real Time Clock Kit £32.00 Assembled £40.00MP -L Parallel Interface Kit £30.00 Assembled £37.00MP -S Serial Interface Kit £30.00 Assembled £37.00PPG -J Joystick Kit £32.00 Assembled £40.00MOD1 UHF Modulator to link CT -64 to T.V. Kit £4.50 Assembled £6.00

PRICES DO NOT INCLUDE V.A.T. or carriage.

SOFTWARESoftware includes 3Kt, 4K and 8K Basic*, CoresidentEditor/Assembler, Text Editing Systemt*, Text Processort*TSC Assemblert*, Games and Scientific packages.* Disk Based versions available.t Full source listing included.Prices vary between £5 and £20.

Both Disk Systems come complete with the comprehensive 'Flex' Disk Operating System. Commands include -

APPEND COPY LINK RENAMEASSIGN DELETE LIST SAVEBACKUP EXECUTE MONITOR STARTUPBUILD GET NEW DISK VERIFYCATALOG JUMP PRINTER VERSION

Applications include: Travel bookings Word Processing Questionnaire analysis Process control Teaching Accounts analysisExisting customers: Schools - Universities - hospitals - M.O.D. - business -industry.

Please write for full details to: -COMPUTER WORKSHOP, 12 TRESHAM ROAD, ORTON, SOUTHGATE, PETERBOROUGH. Tel 073 323 4433 ORCOMPUTER WORKSHOP, 29 HANGING DITCH, MANCHESTER, M4 3ES. 061 832 2269

- BASIC FOR BEGINNERS JV =MI - FOUR EASY PIECES - … · omml JV =MI Jill VOLUME 1, NUMBER 2 50p...

Documents

Transcript of - BASIC FOR BEGINNERS JV =MI - FOUR EASY PIECES - … · omml JV =MI Jill VOLUME 1, NUMBER 2 50p...