Special Issue on Computer Science and BioLife · Windows. Arrow and dot diagrams like this are used...

20
Computer Science for Fun Issue 4 Special Issue on Computer Science and BioLife Kerstin Dautenhahn and her polite robots Win an iPod nano Evolution solution revolution

Transcript of Special Issue on Computer Science and BioLife · Windows. Arrow and dot diagrams like this are used...

Page 1: Special Issue on Computer Science and BioLife · Windows. Arrow and dot diagrams like this are used a lot in computer science to describe how computer systems change over time. Why

Computer Science for Fun Issue 4

Special Issue on

ComputerScience and BioLife

Kerstin Dautenhahn and her polite robots

Win an iPod nano

Evolution solutionrevolution

Pub0319 CS4FN4issue4 v4.qxd 21/11/06 15:12 Page 1

Page 2: Special Issue on Computer Science and BioLife · Windows. Arrow and dot diagrams like this are used a lot in computer science to describe how computer systems change over time. Why

Passionate about computer science?

www.dcs.qmul.ac.uk/cs4fn/

Welcome to cs4fn

In this issue:

Win an iPod:page 9

Politerobots:page 10

Evolutionsolutionrevolution:page 16

2

The FUNdamentals of Computer Science

The trouble with choosing University subjectsis it’s hard to know what the subjects mightbe about. You don’t need to have donecomputer science at school or college to starta CS degree, but how then are you supposedto know what you are getting yourself into?What are those first year topics about? To help, we’ve put together an off-beat cs4fnstyle look at the core topics of ComputerScience. It includes articles, games,

interactive puzzles and video clips groupedby subject theme…It will give you an idea ofthe topics you will study…then when you dogo to University you will have a head starthaving learnt some deep Computer Science.Of course, if you end up doing Law orBiology instead you will still have learnt some useful stuff while having fun.

See www.dcs.qmul.ac.uk/fundamentals/

Not Just a Pretty (!?) Face

Here’s our new cs4fn logo that will bepopping up all over the place in future. If you are wondering what it’s all about, it’s a cartoonized version of a diagram thatfirst appeared on a research poster aboutsomething called Separation Logic, which

is a way of describing computer systems. Seperation logic is being developed atQueen Mary and is causing a lot ofexcitement amongst Computer Scientistsworld-wide including those at Microsoft,where they are building tools around it tohelp remove bugs in programs used withWindows. Arrow and dot diagrams like thisare used a lot in computer science todescribe how computer systems changeover time. Why the happy face? Because,ultimately, computer science is aboutpeople too.

The magazine aboutthe fun side ofcomputer science

When computer science touches othersubjects, they emerge morphed into newand exciting things. Computer scientistsand performance artists meet and outcomes Digital Performance (think Gorillaz –a virtual band that doesn’t exist in the realworld but gives amazing performances).Get biologists or medics and computerscientists talking together and the resultsare new subjects like bioinformatics,biometrics and artificial-life. Amazing newthings are discovered about the way ourbrains work or clues are found in thesearch for cures to cancer. The mostexciting discoveries happen on theboundary of subjects – and computerscience lives on the boundary of lots. In this issue we look at some of the thingsthat have resulted from these interactionsbetween computer scientists and thesquidgy, BioLife subjects. Some reallyexciting stuff emerges out of the gunge.

cs4fn continues to be supported by the CSindustry and the British Computer Society:Intel, Microsoft and ARM have kindlysupported this issue.

Passionate aboutComputer Science

www.dcs.qmul.ac.uk/cs4fn

Pub0319 CS4FN4issue4 v4.qxd 21/11/06 15:12 Page 2

Page 3: Special Issue on Computer Science and BioLife · Windows. Arrow and dot diagrams like this are used a lot in computer science to describe how computer systems change over time. Why

Representing Life … all of it

3

COMPETITION: How do you get

from DNA to Alice in Wonderland?

Above we followed a link from DNA to Gray Code. Alice In Wonderland was written by LewisCaroll and he invented a puzzle that appeared for the first time in Vanity Fair Magazine in1879. Think of two words, like ALICE and GENES. You have to construct a sequence of actualwords starting from the first word and ending with the second. The tricky bit is that no twoconsecutive words in your sequence can differ by more than one letter. ALICE, SLICE, SPICE,… Spotted how that puzzle makes the link to Gray code patented 74 years later?

We have book voucher prizes donated by Intel for the best answer of how to get from ALICE to GENES in this way. The winner’s school will also get voucher prizes. For full details and to submit your answer go to the Magazine+ part of the webzine.

Back at the dawn ofhistory, out of theprimordial soup,something very specialhappened … a codebegan to develop, a codeof life. Things started toreplicate, making copiesof themselves: passing oninformation about howthey were constructed andwhat made them able tosurvive. The code thatdeveloped was a way of representing thatinformation of how to puttogether a life, a creature.A code that allowed allthe information about abody plan to be passed on came into being: themolecule DNA. DNA is a long series of protein molecules.If stretched out, each individual DNAmolecule in my body would be as long as Iam tall. To represent information you needan alphabet. The molecules of life have tobe long because they store a vast amountof information. In English we use an

alphabet of 26 letters to representinformation. Russian has 33 letters. Thealphabet of life has just 4: A, G, C and T(see page 14). Of course the letters usedare rather arbitrary, it is what they meanthat matters. Particular sequences like ACTin particular places in the molecule mightbe part of the instructions that ultimatelyleads to “blue eyes” or “brown hair”. Theseprotein sequences are used in the cell to dothe tasks of life, replicating, building andmaintaining bodies. The language of DNAmakes it easy both for cells to replicate theinformation stored but also to process it inother ways: making functioning creatures. It turns out proteins are a good way to storeinformation. DNA also comes with a wholeseries of mechanisms for detecting errorswhen the information is replicated, and for repairing it.

Computer scientists are also interested in ways of representing information;processing information efficiently is a lot of what the subject is about, and the wayinformation is represented can make a bigdifference. Computers use an alphabet of 2 letters, 1 and 0 (known as bits) torepresent information. One of the thingsthat the alphabet of 1s and 0s is used torepresent are numbers. However there aredifferent ways to do it. For example, thesequence 1101 represents the numberthirteen in the normal way of doing things.Each digit represents a power of 2. The 1s mean add it in, the 0s mean don’t. So1101 is 8+4+0+1. Similarly 0110 means0+4+2+0 or six. This way of codingnumbers makes it very easy to do thingslike multiply and divide. Some other thingsare not so easy though.

Counting 0, 1, 2, 3, 4 in this representationgoes 0000, 0001, 0010, 0011, 0100. To move from one number to the next youneed to change several of the characters at once. If all you want to do is count andnot (say) do addition, then other ways ofrepresenting the numbers might be better.For example, we could make the code zero: 0000, one: 0001, two: 0011, three:0010, four: 0110, etc. This sequence has a special property that counting now onlyinvolves flipping one bit at a time. This kind of code is called a Gray code, after itsinventor, and is used very widely in cableand digital TV systems as it also makes itvery easy to notice when information hasbeen corrupted and to fix it. That correctionof mistakes is just as important inreplicating a TV picture across millions of homes as when replicating a DNAmolecule in your body.

The world revolves around information andboth evolution and computer scientistshave come up with ways of representing it.The ways may be different but the aims ofbeing able to process the information andcatch errors easily are the same.

Try our Medieval ChallengeGo to the cs4fn webzine to see if you are up to a medieval Gray codechallenge: can you crack the lock and open the treasure chest?

Email us at cs4fn�dcs.qmul.ac.uk

Feel free to photocopy pages from cs4fn for personal or class use

Pub0319 CS4FN4issue4 v4.qxd 21/11/06 15:12 Page 3

Page 4: Special Issue on Computer Science and BioLife · Windows. Arrow and dot diagrams like this are used a lot in computer science to describe how computer systems change over time. Why

Passionate about computer science?

www.dcs.qmul.ac.uk/cs4fn/

For years a dream ofcomputer scientists hasbeen to build computersthat can think: one waybeing to try and writeprograms that work likethe brain called NeuralNetworks. Now it turns out that we can help medicsunderstand the brain using the computerscience that people originally thought mightcreate artificial intelligence. Many of theareas of the brain aren't completelyunderstood. Understanding how theseindividual parts work can help in diagnosing diseases and creating cures.

One area of the brain which until recentlywas hardly understood is called the basalganglia. It was thought to be connected todecision making. Different parts of the braincontrol body movements and purely mentaloperations like identifying a smell. Inresponse to the things that happen aroundus ('stimuli'), these brain regions want totake control of our body or our attention.The basal ganglia decides which brain areagets control of the body or the attention,and so what actually happens.

When the brain makes a decision to dosomething because of, say, a noise, youdon't react instantly when you hear thenoise. You only react after a certain amountof stimulus has been received by the brain.It only happens after some threshold hasbeen reached. The basal ganglia wasbelieved to perform this but it wasn’tunderstood how... and the basal gangliacontained some very strange neurons not found in other parts of the brain.

The same type of decision problem is also studied in statistics where there’s aparticular procedure that's the best knownmethod for making decisions on incoming

data. It's described by a complicatedmathematical formula. Bristol Universityresearcher, Rafal Bogacz with otherresearchers at Sheffield, discovered that this formula can be written in a way that issimilar to a neural network: the computerscientist's programmed version of the brain.The formulae can be written as a series ofsimpler parts that pass information betweeneach other. What was surprising was thatthe resulting Neural Network lookedremarkably similar to the wiring that wasfound in the basal ganglia. Could it be thatthe brain had, via the process of evolution,programmed itself to implement a complexstatistical procedure?

So this version of the statistical procedureacts as a model of the brain. Oncescientists have a model of something thenoften they try to predict something thatfollows from the model and see whether this appears in the real world. This givesconfidence that the model is correct. Webelieve Newton's Laws of Motion becausewe can use them to predict where theplanets are in the sky and to predict howobjects travel on earth. Scientific models are a way to see into the future!

A particularly interesting part of the NeuralNetwork model developed was that one part of the model needed to perform acalculation called the exponential function.This is the calculation that converts 0 to 1, 2 to 4, …8 to 256 and so on. In mathsterms if the input, x was greater than zerothen the output would behave like 2x. In other words it would grow very fast.Remember that the basal ganglia had somestrange neurons? Well the model predictsthat the strange neurons should behave likethe exponential function, and when this waschecked they did! The biologists confirmed

that they 'fire' in a way that follows thebehaviour of the exponential function.

So by understanding how to implementstatistics using Neural Networks we nowunderstand parts of the brain better thatwere a mystery. The goal now is to use thismodel to predict what happens when thispart of the brain is damaged. That way we might produce cures for the resultingdiseases. There are still many other aspectsof this area of the brain that we do notunderstand. Maybe we can understandthose in a similar way too.

So we may not have created a computerthat thinks yet, but the computers arehelping us understand how we think!

4

Scientificmodels are away to see intothe future!

Creating a brain inside a computer

Pub0319 CS4FN4issue4 v4.qxd 21/11/06 15:12 Page 4

Page 5: Special Issue on Computer Science and BioLife · Windows. Arrow and dot diagrams like this are used a lot in computer science to describe how computer systems change over time. Why

Biometric Mice

5

In the old days peopleproved their identity by signing their name. Then the world went online and suddenlywe had to remember passwords galore:phrases, numbers, but never a word in adictionary in any language not even Lord of the Rings Elven. Nothing personal thatsomeone could find out about you as that’seasily crackable. Ideally at least 8 letterslong … passwords that are memorable but hard to guess. Impossible! People write them down, tell them to friends andcolleagues, even to random people whophone them up claiming to be from a bank or the IT department. Password gone, secrets gone…and yourmoney…even your whole identity.

Enter biometrics: using some uniquephysical characteristic of the person suchas fingerprints, iris scans or the like. It’sbeen suggested that the biometric industrycould be soon worth $2.5 billion. However,most biometric systems need specialhardware of some kind to measure thecharacteristic. They can also carry a stigmaas well as raising privacy problems. Worse if you can only start your car with afingerprint a grisly consequence might be a robber chopping your finger off to use as

the key. Maybe DNA is the ultimatebiometric ID…but that is easy, if yucky, to steal – after all scientists even extractfragments of DNA from Egyptian mummiesand millennia old insects trapped in Amber.You leave copies of it everywhere you go,flaking skin, saliva on glasses or when yougo to the toilet.

Enter Computer Scientists Peter McOwanand Ross Everitt of Queen Mary. They haveinvented a new way to check your identityonline using a kind of biometrics withoutgrisly or yucky consequences. It couldmake shopping over the Internet muchmore secure and give people moreconfidence buying online. It could even replace signatures and chip and PIN in supermarkets.

The new biometric mouse signature systemuses a written signature, just as with paperbut the person writes it with their mouse.No special gadgets are needed it just usesa standard web browser. The clever part isthat it is not just the pattern on the screenthat matters but the pattern of mousemovements to create it. Sophisticatedartificial intelligence software learns the waythat the person writes the signature andmatches any subsequent writing of thesignature against it. It can also adjust to thegradual change of a signature over time - aproblem with a pencil and paper signature.A fraudulent attempt to sign a documentneeds more than just the pattern itself butto be able to copy the way it was done. Infuture you might not be trying to rememberyet another PIN or password to login, butjust signing your name.

A scientific paper based onresults from the cs4fn onlineSpaceInvaders experimenttogether with a similar cargame experiment just won the best short paper award at HCI2006: a top Internationalresearch conference onHuman-Computer Interaction.Congratulations to Rob Dann and Wai LokCheng, students at Queen Mary on whoseMSci project work the paper was basedand to Jonathan Back at UCL who wrotethe paper. Thanks to all the cs4fn readersfor taking part.

The Space Invaders experiment isinvestigating how hard it is for people to remember to do things in the future. Can you avoid losing points by alwaysremembering to switch on your gun. One of the ways we overcome our imperfectmemories is to manipulate our environment– for example to make sure I remember topost a letter when I go out in the morning, I leave it on the floor in front of the doorwhere I can’t miss it: a message to myselffor the future. Now you can even send anemail to your future self, thanks to a newweb site www.futureme.org. Type themessage as normal but add the date you want it sent.

What do you need to tell your future self?Send the email now before you forget. Then go play spaceinvaders in the webzine.

cs4fnSpaceInvaderswinsresearchaward!

check your identityonline using a kind ofbiometrics withoutgrisly or yuckyconsequences

Email us at cs4fn�dcs.qmul.ac.uk

Feel free to photocopy pages from cs4fn for personal or class use

Pub0319 CS4FN4issue4 v4.qxd 21/11/06 15:12 Page 5

Page 6: Special Issue on Computer Science and BioLife · Windows. Arrow and dot diagrams like this are used a lot in computer science to describe how computer systems change over time. Why

Passionate about computer science?

www.dcs.qmul.ac.uk/cs4fn/6

advantage here is that those images can be processed using clever algorithms tocorrect for problems in exposure, or even to pick up particular shapes in the image.The diagnosis can be helped by the artificialintelligence in the computer, which can spotunusual patterns in the image and alert thedoctor. Better still since these digital X-raysare computer based. They can be easilystored and transmitted throughout the world to places where they are needed.

A slice of life

X-rays, even digital X-rays, can only giveyou flat images of the body innards. Like ashadow they squash all the depth details.Your insides are three-dimensional (3D)though so it would be useful to be able to slice through your body and get a viewinside. This is possible using a computerbased method called tomography, from the Greek tomos (slice) and graphia(describing). It still uses X-rays but intomography the X-ray source and thedetector rotate round the body taking lots ofimages at different angles. It’s like castingdifferent shadows as the sun moves round

Computer scientists are helping doctors,surgeons, biologists andpsychologists get insidethe body and mind, andimproving the way thatmedical care will beprovided now and in thefuture. It’s a fascinatingstory of biology, mathsand computing and it all starts with an X.

What a picture!

X-rays were the first practical method ofexamining the inside of a living body. Theprocess involves firing high energy X-raysthrough the body with a photographic plateat the other side. Dense bits of the bodylike bones absorb radiation. That leads to alighter area on the developed photographicnegative. In effect a shadow is cast throughyou onto the photograph, giving a viewinside. A problem with this is that, as withany camera, it’s hard to get the photographexposure right. Worse you have to find thespace to store hundreds and thousands ofsheets of film. Worse still, suppose yourdoctor in Manchester needs the X-ray takenof you when you are wanting to play footballso you are in Frankfurt. The film has to besent by post. Enter computer scientists tomake things easier.

Portable pixelpictures

New digital X-ray systems are beingdeveloped. These use X-ray detectors notfilm and produce digital images rather thanthe standard photographic images. The

Looking insidemedicine-computer scientists in the body

Pub0319 CS4FN4issue4 v4.qxd 21/11/06 15:12 Page 6

Page 7: Special Issue on Computer Science and BioLife · Windows. Arrow and dot diagrams like this are used a lot in computer science to describe how computer systems change over time. Why

you. So imagine you’re using tomographyon a cylinder, and your X-ray source is atorch. Move the torch round the cylinderand look at the shadow cast on a piece ofpaper moving at the opposite side to thetorch. Each ‘shadow’ picture would look the same because a cylinder is circularlysymmetric. Now imagine a more interestingshape. Each of the shadow pictures woulddepend on where you were at the time inrelation to the shape. With some clevermaths, a reconstruction algorithm and acomputer you can go from the shadowpictures back to the shape. These shapesare the organs and innards of your body,and they can be recorded in their full 3Dglory. There are now systems that spiral theX-ray source round the body making itquicker. You can even do tomography atvery high speed allowing slices through thebeating heart to be calculated. Interestinglythe maths behind this technology, calledthe “Radon transform” after Czechmathematician Johann Radon(1887–1956), was developed purely as an abstract mathematical theory. No one at the time could see any use for it!

Go to the Magazine+

section of the cs4fn

webzine for more on

how magnets and

wobbly protons mean

we can now even see

inside your squishy bits.

Email us at cs4fn�dcs.qmul.ac.uk

Feel free to photocopy pages from cs4fn for personal or class use 7

Check in at theDigital Hospital

Life saving healthcare and medical imagingis going digital. Using video conferencing,mobile scanners and even remote operatedrobotic surgery the field of tele-medicineallows expert medical care to be providedany time, any place. Today’s progresstowards the digital hospital combinesdifferent ways of taking information aboutthe state of your body, such as digital X-rays, or tomographic images, readingsfrom digital thermometers or digital bloodpressure readers. We can combine all thisinformation with your personal informationinto one big file, so there is no need formultiple paper copies to get out of date orlost. The hospital information system keepstrack of all your data, and also importantlywho has access to it.

Tomorrow’s

world and you

According to Alan McBride, a computerscientist who is working on these state of the art medical systems:

‘This technology is a major step forward in health care where the UK is leading the way. The government’s grand scheme will allow images taken in Newcastle to be shown on your GP’s desk in London,together with the hospital report, which will automatically be emailed to their inbox.Computer science is playing the major rolein all this, creating new ways to aid clinicalpractice, with plenty of scope in the futurefor talented computer scientists to getinvolved.’

The computer scientists who make thishappen will not only be technical specialistsbut also experts in understanding humanbehaviour. We will only get the benefitssuch a grand scheme promises if theconflicting needs and concerns of all those involved are taken into account:patients, nurses, doctors, managers and politicians…that will take major people skills.

Pub0319 CS4FN4issue4 v4.qxd 21/11/06 15:13 Page 7

Page 8: Special Issue on Computer Science and BioLife · Windows. Arrow and dot diagrams like this are used a lot in computer science to describe how computer systems change over time. Why

The mostexcitingresearch oftencomes aboutwhen peoplefrom differentsubjects worktogether and itcan come fromunexpecteddirections. Howmight ComputerScientists help inthe discovery ofdrugs to fightCancer?

The obvious way is in building faster super-computers to number-crunch the problems.However, there are more subtle ways. Ideasfrom one area can have a big impact onothers, if the researchers are creativeenough to see the links. Professor MuffyCalder, a Computer Scientist at GlasgowUniversity, discovered one intriguing linkwhile working with cancer specialists. It turns out that the problems ofunderstanding how drugs act on our cell chemistry are very similar to those inunderstanding communication networks.Tools for one can be used for the other.

Muffy and team are aiming to understandthe biochemical “pathways” in whichsignals pass through from the membrane of cells into their nucleus. These“pathways” are just a series of chemicalreactions where different protein moleculesare created and destroyed. To developcancer drugs, scientists need to have abetter understanding of how this happens.It will help them predict the way thereactions are affected by drugs.

The pathways are normally modelled usingcomplex maths built into tools that simulatethe processes involved – allowing virtualexperiments to be done on the computerinstead of on real cells. These simulationsare used to suggest actual experiments andto help understand the results. Muffy’steam realised that the diagrams used bybiochemists to illustrate the reactions arewell known to computer scientists – theylook just like the “producer-consumer”

Passionate about computer science?

www.dcs.qmul.ac.uk/cs4fn/8

Predicting Cancer Cures

networks already used to analyse telecomnetworks. That means tools alreadydeveloped for analysing telecom networkscan be used to analyse the biologicalnetworks. What advantage does this give?Proof can be used rather than justsimulation. Simulation allows you to checkwhat happens in particular individualsituations: the ones you simulate.Unfortunately other situations may or maynot give the same results: you don’t knowunless you simulate them too. With themathematical proof-based tools, generalproperties of the biological system asmodelled can be shown to always be true. You can, for example, work out theprobability that too much of a particularprotein will be produced.

Previous work like this had a focus onproving properties at a molecular level:what will happen if single molecules react?Muffy realised that when developing drugsit’s not what single molecules do that

ultimately matters, but being able to predicthow test tubes of reagents (substances thatreact) behave. That means you have tomodel something slightly different: what’sknown as the molar level, the test-tubelevel. It turns out that to prove propertiesinteresting to the biochemists only twolevels of concentration of reagents matter -high and low - corresponding to enough

and not enough reagent to trigger areaction. This is a similar idea to onecomputer scientists use to reason abouthardware circuits, thinking in terms of highand low voltage levels rather than aboutindividual electrons.

The work opens up a whole new approachfor developing drugs. The ultimate aim is to provide predictive tools for biochemists.They will suggest what effects differentdrugs might have on the processes takingplace in cells and so suggest experimentsto perform. If the team of computerscientists and cancer specialists domanage this, they will have handed the biochemists a powerful new tool to help in the fight against cancer.

Pub0319 CS4FN4issue4 v4.qxd 21/11/06 15:13 Page 8

Page 9: Special Issue on Computer Science and BioLife · Windows. Arrow and dot diagrams like this are used a lot in computer science to describe how computer systems change over time. Why

Biology and Computer Sciencecan meet in some unexpected,not to mention inhospitable,places. Who would have thoughtthat the chemical soup in thenests of Stormy Petrels studiedby field biologists might help in the development of futuristiccomputer technologies at Intel,for example.

Just Keep DoublingOne of the most successful predictions inComputer Science was made by GordonMoore, co-founder of Intel. Back in 1965he suggested that the number of transistorsthat can be squeezed onto an integratedcircuit – the hardware computer processorsare made of - doubled every few years:computers get ever more powerful and ever smaller. In the 40 or so years sinceMoore's research paper it has remained an amazingly accurate prediction. Will itcontinue to hold though or are we reachingsome fundamental limit? Researchers atIntel are confident that Moore's Law can berelied on for the foreseeable future. Thechallenge will be met by the MaterialScientists, the Physicists and the Chemists.Computer Scientists must then be ready for the Law’s challenge too: delivering thesoftware advances so that its trendsare translated into changes inour everyday lives. It will

lead to ever more complex systems on asingle chip and so ever smaller computersthat will truly disappear into theenvironment.

Dusting computersMotes are one technology underdevelopment on the back of this trend.They are dust-sized computers. Scatterthem around the environment and theyform unobservable webs of intelligentsensors. Scatter them on a battlefield todetect troop movements or on roads tomonitor traffic flow. Mix them in concreteand monitor the state of a bridge. Embedthem in the home to support the elderly or in toys to interact with the kids.

What barriers must be overcome to makeMotes a ubiquitous reality? Much of thearea of a computer is taken up by itsconnections to the outside world - all those pins allowing things to be plugged in. They will be replaced by wirelesscommunications. At the momentcomputers contain multiple chips eachhousing separate processors. It is not thecircuits that are the problem but thepackaging - the chip casings are both bulkyand expensive. The future is for what isknown as “multicore” chips: large numbers

Email us at cs4fn�dcs.qmul.ac.uk

Feel free to photocopy pages from cs4fn for personal or class use 9

The Tale of theStormy Petrel

of processors on a single small chipcourtesy of Moore's Law. This will lead to significant challenges to ComputerScientists over how to develop the softwareto run on such complicated hardware.Battery technology is the only big problemthat remains. Motes will soon be with us.

Enter the PetrelsBut how do you test a device like that?Enter the Stormy Petrel birds. Intel’sapproach is not to test futuristic technologyon average users but to look for extremeones who believe a technology will deliverthem massive benefits. In the case ofMotes, the extreme users are field biologistswho want to keep tabs on birds inextremely harsh field conditions. Not only isit physically difficult for humans to observesea birds’ nests on inhospitable cliffs buthuman presence disturbs the birds. Thesolution: scatter motes in the nests todetect heat, humidity and the like fromwhich the state and behaviour of the birdscan be deduced. A nest is an extremelyharsh environment for a computer though,both physically and chemically. A wholebunch of significant problems, overlookedby normal lab testing, must be overcome.The challenge of deploying Motes in such a harsh environment has led to majorimprovements in the technology.

Moore's Law is with us for a while yet, andwith the efforts of Material Scientists,Physicists, Chemists, Computer Scientistsand even Field Biologists and their seabirds it will continue to revolutionise ourlives.

Using GeomLab, developed by Mike Spivey of theUniversity of Oxford with support from NAGTY, youmust program a picture. Simple pictures comeeasy but even ones as complicated as Escher'sbeautiful, thought-provoking pictures with ever-repeating lizards, fish and other animals that you may have come across, are possible.

GeomLab is a great way to learn about this different style of programming calledfunctional programming. You will also get an idea of the issues in laying out the elements of web pages.

See the Magazine + section ofthe webzine for full detailsand how to enter.

Win an ipod nano

Win great prizesfor you and yourschool ...andlearn a fun way toprogram pictures

iPod nano andbook vouchers upfor grabs. Prizesprovided by ARM

and Intel

Pub0319 CS4FN4issue4 v4.qxd 21/11/06 15:13 Page 9

Page 10: Special Issue on Computer Science and BioLife · Windows. Arrow and dot diagrams like this are used a lot in computer science to describe how computer systems change over time. Why

Passionate about computer science?

www.dcs.qmul.ac.uk/cs4fn/10

KerstinDautenhahn is abiologist with amission: to helpus make friendswith robots.

Kerstin was always fascinated by thenatural world around her, so it was nosurprise when she chose to study Biologyat the University of Bielefeld in Germany.Afterwards she went on to study a Diplomain Biology where she researched into theleg reflexes in stick insects, a strange startit may seem for someone who would laterbecome one of the world’s foremostrobotics researchers. But it was throughthis fascinating bit of biology that Kerstinbecame interested in the ways that livingthings process information and control theirbody movements, an area scientists callbiological cybernetics. This interest in tryingto understand biology made her want tobuild “things” to test her understanding.These “things” would be based on ideascopied from biological animals but be run bycomputers. These “things” would be robots.

Follow thatrobot

From humble beginnings building smallrobots that followed one another over a hilly landscape, she started to realise thatbiology was a great source of ideas forrobotics, and in particular that the socialintelligence that animals use to live andwork with each other could be modelledand used to create sociable robots. Shestarted to ask fascinating questions like"What's the best way for a robot to interruptyou if you are reading a newspaper - bygesturing with its arms, blinking its lights or

making a sound?" and perhaps mostimportantly “When would a robot becomeyour friend?” Now at the School ofComputer Science at the University ofHertfordshire where she is a Professor of Artificial Intelligence, she leads a worldfamous research group looking to try andbuild friendly robots with social intelligence.

Good robot /Bad robot –East vs West

Kerstin, like many other roboticsresearchers, is worried that most peopletend to look on robots as being potentiallyevil. If we look at the way robots areportrayed in the movies that’s often how it seems: it makes a good story to have amechanical baddie. But in reality robotscan provide a real service to humans, fromhelping the disabled, assisting around thehome and even becoming friends andcompanions.

The baddie robot ideas tend to dominate in the west, but in Japan robots are verypopular and robotics research is advancingat a phenomenal rate. There has been along history in Japan of people findingmechanical things that mimic naturalthings interesting and attractive. It is partlythis cultural difference that has madeJapan a world leader in robot research. But Kerstin and others like her are trying to get those of us in the west to change our opinions by building friendly robots and looking at how we relate to them.

Future FriendlyFocus on Kersti

This fall inacceptability

is called the‘uncanny

valley’

Pub0319 CS4FN4issue4 v4.qxd 21/11/06 15:13 Page 10

Page 11: Special Issue on Computer Science and BioLife · Windows. Arrow and dot diagrams like this are used a lot in computer science to describe how computer systems change over time. Why

Email us at cs4fn�dcs.qmul.ac.uk

Feel free to photocopy pages from cs4fn for personal or class use 11

Polite robotsroam the room

Kerstin decided that the best way to seehow people would react to a robot aroundthe house was to rent a flat near theuniversity, and fill it with robots. Ratherthan examine how people interactedwith robots in a laboratory, moving the

experiments to a home, with bookcases,biscuits, sofas and coffee tables, make itreal. She and her team looked at how togive their robots social skills, what was thebest way for a robot to approach a person?At first they thought that the best approachwould be straight from the front, but theyfound that humans felt this too aggressive,so the robots were trained to come upgently from the side. The people in thehouse were also given special ‘comfortbuttons’, devices that let them indicate howthey were feeling in the company of robots.Again interesting things happened, itturned out that not all, but quite a lot ofpeople were on the whole happy for theserobots to be close to them, closer in factthan they would normally let a humanapproach. Kerstin explains ‘This is becausethese people see the robot as a machine,not a person, and so are happy to be inclose proximity. You are happy to moveclose to your microwave, and it’s the samefor robots’. These are exciting first steps aswe start to understand how to build robotswith socially acceptable manners. But itturns out that robots need to have goodlooks as well as good manners if they aregoing to make it in human society.

Looks areeverything for a robot?

How we interact with robots also dependson how the robots look. Researchers hadfound previously that if you make a robotlook too much like a human being, peopleexpect it to be a human being, with all thesocial and other skills that humans have. If it doesn’t have these, we find interaction

ly: in Dautenhahn

Kerstin and her team

Pub0319 CS4FN4issue4 v4.qxd 21/11/06 15:13 Page 11

Page 12: Special Issue on Computer Science and BioLife · Windows. Arrow and dot diagrams like this are used a lot in computer science to describe how computer systems change over time. Why

12

very hard. It’s like working with a zombie,and it can be very frightening. This fall inacceptability of robots that look like, butaren’t quite, human is what researchers call the ‘uncanny valley’. People prefer toencounter a robot that looks like a robotand acts like a robot. Kerstin’s group foundthis effect too, so they designed their robotsto look and act the way we would expectrobots to look and act, and things got muchmore sociable. But they are still looking athow we act with more human like robotsand have built KASPAR, a robot toddler,which has a very realistic rubber face

capable of showing expressions andsmiling, and video camera eyes that allowthe robot to react to your behaviours. Hepossesses arms so can wave goodbye orgreet you with a friendly gesture. He’s very lifelike and hopefully as KASPAR’sprogramming grows, and his abilitiesimprove he will emerge from the uncannyvalley to become someone’s friend.

Autism – mindblindness androbots

The fact that most robots at present looklike and act like robots can help insupporting children with autism. Autism is a condition that prevents you fromdeveloping an understanding of how tointeract socially with the world. A currenttheory to explain the condition is that thosewho are autistic cannot form a correctunderstanding of others’ intentions. It’scalled mind blindness. For example if Icame into the room wearing a hideous hatand asked you ‘Do you like my lovely newhat?’ you would probably think, ‘I don’t likethe hat, but he does, so I should say I likeit so as not to hurt his feelings’. You have a mental model of my state of mind (that Ilike my hat). An autistic person is likely torespond ‘I don’t like your hat’, if this is whathe feels. Autistic people cannot create thismental model so find it hard to makefriends and generally interact with people,as they can’t predict what people are likelyto say, do or expect.

Playing withRobot toys

It’s different with robots: many autisticchildren have an affinity with robots. Robots don’t do unexpected things. Theirbehaviour is much simpler, because theyact like robots. Using robots Kerstin’s grouphave been examining how we can use thisinteraction with robot toys to help someautistic children to develop skills to allowthem to interact better with other people.By controlling the robot’s behaviours someof the children can develop ways to mimicsocial skills, which may ultimately improvetheir quality of life. There is no finalconclusion yet, but some promising results,and this work continues to be one way totry and help those suffering with thissocially isolating condition.

Futurefriendly

It’s only polite that the last word goes to Kerstin

‘I firmly believe that robots asassistants can potentially bevery useful in many applicationareas. For me as a researcher,working in the field of human-robot interaction is excitingand great fun. In our team we have people from variousdisciplines working togetheron a daily basis, includingcomputer scientists, engineersand psychologists. Thiscollaboration, where peopleneed to have an open mindtowards other fields, as well as imagination and creativity,is necessary in order to makerobots more social.’

In the future, when robots become ourworkmates, colleagues and companionsit will be in part down to Kerstin and herteams pioneering effort as they worktowards making them future friendly.

Passionate about computer science?

www.dcs.qmul.ac.uk/cs4fn/

Pub0319 CS4FN4issue4 v4.qxd 21/11/06 15:13 Page 12

Page 13: Special Issue on Computer Science and BioLife · Windows. Arrow and dot diagrams like this are used a lot in computer science to describe how computer systems change over time. Why

13

Communicatingwith computersis clunky to saythe least - weeven have to goto IT classes tolearn how to talk to them. It would be somuch easier ifthey went toschool to learnhow to talk to us.

If computers are to communicate morenaturally with us we need to understandmore about how humans interact with each other. The most obvious way that wecommunicate is through speech – we talk,we listen – but actually our communicationis far more subtle than that. People pick up

lots of information about our emotions andwhat we really mean from the expressionsand the tone of our voice – not from whatwe actually say. Zabir, a student at QueenMary who went on to work for a MerchantBank was interested in this so decided toexperiment with these ideas for his finalyear project. He used a kit called LegoMindstorm that makes it really easy to build simple robots. The clever stuffcomes in because, once built, Mindstormcreations can be programmed withbehaviour. The result was Blade.

Blade, named after the Wesley Snipes film,was a robotic face capable of expressingemotion and responding to the tone of theuser’s voice. Shout at Blade and he wouldlook sad. Talk softly and, even though hecould not understand a word of what yousaid he would start to appear happy again.Why, because your tone says what youreally mean whatever the words – that’swhy parents talk gobbledegook softly tobabies to calm them.

Blade was programmed using a neuralnetwork, a computer science model of theway the brain works, so he had a brainsimilar to ours in some simple ways. Blade learnt how to express emotions verymuch like children learn – by tuning theconnections (his neurons) based on hisexperience. Zabir spent a lot of timeshouting and talking softly to Blade,teaching him what the tone of his voicemeant and so how to react. Blade’sbehaviour wasn’t directly programmed, it was the ability to learn that wasprogrammed.

Eventually we had to take Blade apartwhich was surprisingly sad. He really didseem to be more than a bunch of legobricks. Something about his very humanlike expressions pulled on our emotions:the same trick that cartoonists pull with thebig eyes of characters they want us to love.

Watch the video of Blade in

action in the cs4fn webzine.

the emotional computer

Hans Moravec,

of Carnegie

Mellon

University,

has suggested

it may be

possible to

download the

information in

your brain and

store it in a

digital format

as a way to

achieve

immortality…

others aren’t

so sure…

Email us at cs4fn�dcs.qmul.ac.uk

Feel free to photocopy pages from cs4fn for personal or class use

Pub0319 CS4FN4issue4 v4.qxd 21/11/06 15:13 Page 13

Page 14: Special Issue on Computer Science and BioLife · Windows. Arrow and dot diagrams like this are used a lot in computer science to describe how computer systems change over time. Why

Passionate about computer science?

www.dcs.qmul.ac.uk/cs4fn/14

Combiningmusical notesmakes all themusic in theworld; fromMozart to theArtic Monkeysit’s simply a caseof creativelyputting theright notestogether tomake a tune.

A musical score, is a sequence of markstelling you what to play next. It’s a way torecord the instructions. All life on earth hasits own set of instructions, but rather thanbeing marks on a paper to tell you how toplay the tune, life uses chemicals in a longthin molecule called DNA to code theinstructions for building living things.

Give me an A - C - G - T

DNA, or Deoxyribonucleic acid, is a veryclever molecule. It needs to be. After all its storing all those important life buildinginstructions. Its actually two long strings ofchemicals that wrap round each other in aspecial way, called a double helix. It’s a bitlike two spiral staircases twisted together,and it lets the molecule store the

instructions twice, for safety. Theinstructions on the individual strands are made up of chemicals called bases.There are four of them, called adenine(abbreviated A), cytosine (C), guanine (G)and thymine (T), and its these four lettersthat make you what you are. When the DNA is active each pattern of these bases(normally thought of in groups of 3 basese,g, AAC and called a codon) allows a cellto put together a chunk of a protein in the correct order. Following the codoninstructions on the DNA, the cell can buildlong complex protein chains. These chainsthen float off into the cell and chemicalforces within the chains cause them to twistinto particular shapes. It’s the shape ofproteins that let them do their work. Theyact like a lock and key, fitting their threedimensional shape onto the surface of otherproteins and chemicals and so allowingchemical reactions to happen.

Genetics andgarden genomes

All life on earth has DNA as its basis. It’s theway that information on what it means to bea particular plant, flower, virus or animalpasses from one generation to the next,along with instructions on how to buildthem. The DNA is the data store of life;each species has its own set of DNA, oftenbundled into larger structures calledchromosomes. This full set of data to builda species is called the species genome, andunderstanding how it works is one of thegreatest challenges in current science.

Can you see apattern here?

Biology and computer science have oftenbeen seen as separate. Biology looks atplants and animals, computer science atprograms. But when you start to look at thegenetic basis of life, its all about patterns of data. With DNA it’s the codons with theirfour letter alphabet, while in computerscience it’s tended to be the two letteralphabet of binary. Computer scientists can cope with some extra letters, and

today the fascinating and expanding field of bioinformatics takes the techniques ofcomputer science and applies them to helpuntangle the world of biological genetics.

A pattern of guilt

How do we know that, for example, aparticular blood sample found at a scene of a crime belongs to the suspect? We cantake the sample, and using some cleverchemistry amplify the amount of DNA in it to amounts that we can measure moreeasily. We then look at the pattern of theDNA from the crime scene and compare it to the pattern of DNA taken from thesuspect. If the patterns match then youhave evidence of guilt. DNA is like afingerprint: each person has a uniquepattern. A computer vision system scans thepatterns and tell us if there is a match. Thisability of computers to rapidly find whethertwo patterns are the same is at the core ofwork in genetics to try and decode thegenome. Let’s suppose that a particularsequence codes up a particular protein inone species. Imagine we then find thatpattern in another new species’ DNA. Wehave strong evidence that the new speciesalso produces that particular protein, andthat helps us understand how the newspecies lives.

It’s a stringthing

Long lists of text or numerical elements,such as AGGTAC or 122552424, are calledstrings in computer science. There is a longhistory of computer scientists inventingclever ways to efficiently find stringmatches. It’s important to them. Forexample when you type in your computerpassword the password string is comparedwith the stored password string. If thesestrings match exactly you are grantedaccess. So we can take these techniquesover to use in biology where it’s exactly the same problem we need to solve, just different types of strings.

Music in thekey of life

Pub0319 CS4FN4issue4 v4.qxd 21/11/06 15:13 Page 14

Page 15: Special Issue on Computer Science and BioLife · Windows. Arrow and dot diagrams like this are used a lot in computer science to describe how computer systems change over time. Why

15

Bioinformatics:It’s a BLAST

BLAST which stands for Basic LocalAlignment Search Tool, is a computeralgorithm that quickly searches for matchingpatterns in genetic data. BLAST searchcompares the genetic data you want toexamine (often called the query sequence)with a library or database of previouslyidentified sequences. However, often with

biological data no two sequences are exactlythe same, term by term, so BLAST gives anestimate of how well the sequences matchto those in the library. To start with BLASTtries to get a perfect match. The algorithmdefines a range, or window, over which itwill search for a good match. For example,if the algorithm was given the sequencesAGGTAC and ACGTAG and a window size of 3, it would pick out the matching three-element pattern GTA that is in bothsequences. For real data the window size is normally larger, typically looking formatches of a dozen or so elements.

Let’s try thatagain

BLAST has identified a part of the sequencethat’s an exact match between the querystring and the library string, but there’smore to do. In the second stage, BLASTtries to extend the match in both directions,starting at the point where it found the exactmatch. It looks for a match at each letter asit moves out, adding 1 to its score value ifthe letters match, and giving no score ifthey don’t. In this way you can extend thesearch. A low score tells you the matchingisn’t good, but it may be good enough togive you some useful information. There are

The music of life

DNA is the musical score for life, and computer scientists are helping us to understand its beauty and complexity. Music tends to repeatphrases, as does DNA. The same pattern of notes are played in a tune. The same patterns of codons occur again and again in a genome.In fact both music and DNA share another thing in common. They both have a similar distribution of repetitions and it crops up all overthe natural world. Intrigued by this fact many people have actually used DNA sequences to produce music! Some say that the music ofDNA sounds a bit like Mozart, but whatever tune DNA plays, understanding how it works will be top of the science charts for many, manyyears to come.

of course lots of improvements that can andhave been made to this algorithm. They letyou take into account any gaps in thepatterns, but the basic idea is the same,and this method is employed all over theworld to search and help understand themassive amount of genomic data that hasbeen amassed over the last few years.

DNA and theProsecutor’sFallacy

DNA fingerprint evidence can close acase. If you have found a DNA matchthe result is certain…or is it? Why did the Society for Expert Witnesses call in aQueen Mary Computer Scientist to makea video for them explaining clearly whatyou can deduce from such “fingerprint”evidence? …Because lawyers often get it wrong even in high profile cases,claiming cast iron cases when they havenone at all. Turns out tools developed for Software Engineers help.

Go to the webzine tofind out more…

Email us at cs4fn�dcs.qmul.ac.uk

Feel free to photocopy pages from cs4fn for personal or class use

A Slice of Life

The Visible Human Project aims to createa complete, anatomically detailed, 3Dmodel of male and female human bodies.Using computer assisted imagingmethods like tomography this gigantic set of images are available online forresearchers to examine and addinformation to. A grizzly aspect is that thebodies used for the project were actuallycadavers (dead bodies), which were cutup into thousands of thin slices. The malebody was sliced at one-millimetreintervals, the female at one-third-of-a-millimetre intervals, and the data digitisedfor the project. Both had left their bodiesto science – the man was actually acriminal who had been executed.

Pub0319 CS4FN4issue4 v4.qxd 21/11/06 15:13 Page 15

Page 16: Special Issue on Computer Science and BioLife · Windows. Arrow and dot diagrams like this are used a lot in computer science to describe how computer systems change over time. Why

Passionate about computer science?

www.dcs.qmul.ac.uk/cs4fn/16

from single light detecting cells to completeeyeballs that each work in a way that wouldhelp a creature survive. This compares wellin fact with the range of similar “eyes”found in real creatures.

Evolvinganswers

In the other direction the ideas of evolutionhave given computer scientists ideas fornew ways to program – the notion of geneticalgorithms. Computer science is aboutcoming up with solutions to problems andthat is exactly what nature does over time –adapt animal species via natural selection to allow them to survive better in theirchanging environments. The idea is that to solve a problem you turn it into “digitalDNA” and evolve a solution.

Instructive DNA

DNA is the biological molecule that codesthe instructions for creating all living thingson earth. It’s a sequence of instructions,data, that cells use to build new cells, andit’s also the key to evolution. When speciesmate, part of the mother’s and part of thefather’s DNA combine to give the newinstructions for the child. That’s why youlook a bit like your mum and dad. Parts oftheir data were used to build you. That DNAhas been crossed over. When a species isunder environmental threat often onlyparticular types of offspring survive: thosethat can run fastest, or climb highest orblend into the background best. Inspired by Darwin, this is what English PhilosopherHerbert Spencer pithily called “survival ofthe fittest”; the new set of instructions in anoffspring, combined from parents, can givethat offspring a survival advantage. Survivalmeans this creature can go on to breeditself and will pass those useful instructionsto its offspring and so on. Very occasionallydue to the effects of cosmic rays or otherfactors the DNA can be changed, the

In theory itshould explainsomething

Good scientific theories are predictive.They don’t just describe the way theworld is as we know it but suggestexperiments or other investigation thatwe don’t know the answer to. If thetheory is correct then the predictionsshould turn out to be true. The theoryshould also be able to explain futureobservations. That’s why Darwin’s theoryof evolution by natural selection isaccepted by scientists. It’s sometimesargued that evolutionary theory can’t be good science because you can’t doexperiments about the past. There’smore to scientific investigation than testtube science though. There have beenvastly many correct predictions based onthe theory of natural selection (see Box).The most profound is that it predictedthe need for a biological way for passinginformation between generations. Thatultimately led to the discovery of theDNA molecule and within half a centurythe mapping of the human genome aswell as that of other animals. Where doescomputer science come in? Well it worksboth ways.

ComputerSciencenaturally

Computer simulations can be used toexplore whether things might bepossible. For example, some people findit hard to believe that complex biologicalfeatures like an eyeball could possiblycome about step by step as predicted bynatural selection. Computer simulationshave shown that such things arefeasible: for example an early computerexperiment showed it is possible toconstruct sequences of partial eyeballs

The evolutionsolutionrevolution

The evolutionsolutionrevolution

Pub0319 CS4FN4issue4 v4.qxd 21/11/06 15:13 Page 16

Page 17: Special Issue on Computer Science and BioLife · Windows. Arrow and dot diagrams like this are used a lot in computer science to describe how computer systems change over time. Why

Email us at cs4fn�dcs.qmul.ac.uk

Feel free to photocopy pages from cs4fn for personal or class use 17

chemical sequences are mutated. Normallythis causes death. The instructions arecorrupted and the offspring dies, but fromtime to time a non-lethal mutation occurs,and this can lead to offspring with a newsurvival trick. If it’s advantageous it will letthe offspring and its descendants thrive. In effect the process of evolution keeps thespecies and the environment matched toone another. You could say the speciessolves the problem of survival in theenvironment.

Visiting with a GeneticAlgorithm

Computer scientists can now take aproblem, code it up as a sort of digital DNAand use the evolutionary model, crossoverbetween parents to create offspring andmutation, to find a good solution. A classicexample is the Travelling SalespersonProblem. In this problem you have to visitsay twenty towns, and do the least travelling(you need to keep the travelling expensesdown). We start by randomly creating apopulation of ‘creatures’ whose digital DNAis simply a list of the twenty towns (saynumbered 1-20) initially in random order. We then test how fit each of these‘creatures’ is by calculating the amount oftravel expenses each of them requires (howexpensive that route would be to follow).From our initial population of possiblesolutions we choose the best solutions,those with the lowest travel expenses, andlet these breed together: making a newroute with parts from each parent. Thosesolutions with high costs don’t survive in thislow cost financial environment, and aresimply killed off, removed from the program.The next generation, made by swapping(crossover) the digital DNA of the originalbest solutions, are then tested, again thebest solutions survive and we breed them tocreate a new generation and so on. Manythousands of generations later we havereally good solutions, which have survivedbecause they are a good solution to theproblem.

Mutate and explore

We always have a small random mutationrate in the background, as this parentoffspring breeding can often result in youonly exploring a good solution, they arebreeding in only one isolated part of all the possible good solutions, and possiblymissing the best solution. That may requiresomething very different.

After running a genetic algorithm on yourproblem you simply turn the digital DNAback into, in this case the list of towns inorder, and you’re done. Problem solved.

Evolutionworks

There are many different variants on thebasic genetic algorithm which try to ensurethat they find the best solution possiblerather than just a good solution. This abilityto find the very best, or global solution,cannot yet be guaranteed mathematically –nature doesn’t guarantee perfection either.However genetic algorithms are useful andpowerful and have been used to solveproblems in water distribution andcomputer networks, robotic learning, schooltimetabling and even back into biology andchemistry to calculate the shapes of foldedproteins or atomic forces.

So computer science gives a way ofexploring predictions about the details of how evolution works and can even helpsuggest things to look for in nature. On theother hand, by understanding how nature,through DNA and evolution leads to goodsolutions to ensure survival in oftenincredibly harsh environments, scientistshave created a powerful new form ofcomputing in the genetic algorithm.

Try it out – join a sodarace(www.sodarace.net) and see how powerfulgenetic algorithms can tune a virtual beastto its current environment!

Fossils for the record

Recently palaeontologists went looking forfossils on Ellesmere Island in the Arctic.They were specifically interested infinding fossils that made the link betweenfish and land living animals. Why did theylook there? Not because fossils had beenfound there before but because theyknew from existing fossils from elsewherethe time, 400 million years ago, whensuch animals should have existed. Nextthey used their knowledge of geographyto work out where rocks of that ageexisted. Ellesmere Island fitted so that is where they looked. After 5 years ofdigging on the island, earlier this year,they found near perfect skeletons of fishwith wrist like arrangements that onewould expect from the missing link.

Beak to thefuture

A Herculean study showing the amazingdedication of some scientists is that ofRosemary and Peter Grant. They did fieldwork on small remote islands in theGalapagos documenting evolution as ittook place in the islands’ finches. Theydid this over a staggering 30 years. Theycarefully measured the beaks of thefinches generation after generation, alsorecording things like the weather andkinds and quantities of seeds availableeach year. Based on the data theycollected they were able to predict whatwould happen to beak sizes whendroughts happened.

EvolutionaryPredictions

Pub0319 CS4FN4issue4 v4.qxd 21/11/06 15:13 Page 17

Page 18: Special Issue on Computer Science and BioLife · Windows. Arrow and dot diagrams like this are used a lot in computer science to describe how computer systems change over time. Why

Passionate about computer science?

www.dcs.qmul.ac.uk/cs4fn/18

What do thebook The DaVinci Code, TV show BigBrother,computersoftware andcomputer chipdesigns have incommon? Theyare all based onideas and allform somethingcalledintellectualproperty.

It can be a big issue for the computerindustry. Computer firms like ARM sell ideas not tangible chips. They designmicroprocessors, in effect licensing theirideas to other companies who actuallycreate the microprocessors and put them in their mobile phones or game consoles.Are programs just ideas? After all you can’ttouch the code. It’s just bits. So what doesthe law protect? Will someone else maketheir fortune from your great ideas?

It’s mine, Ithought of it first

If you were to craft a lovely, smooth woodentable, and someone came along and nickedit the law would be on your side. Your tableis physical property. You can polish it, youcan lend it to a friend for a party or you cansell it. This kind of protection for physicalobjects has been enshrined in law forcenturies. However in today’s informationeconomy it’s often ideas, rather thancarpentry, that are the key to making loadsof money. So how do you go aboutprotecting the mental, or intellectual, effortyou put into thinking up say a new TV show,a plot for a book, a movie or a mathematical

algorithm in a computer program, whenthere is nothing physical about it? You can’ttouch it, but it’s still something worthwhileyou have created.

Wanna buy myscribble?

The law protects intellectual property, thingsbuilt with your brain rather than your hands,in several ways. One of the main forms ofprotection is copyright. Copyright is a lawthat goes back centuries. It gives the creatorof a work the right to decide how the workcan be used. Copyright is automatic,scribble on a bit of paper, and so long as it’s permanently recorded in some way it’scopyrighted. You have the power to decidewho can use your doodle, or make copies ofit, and you can sell this right on if you canfind someone willing to buy. What’s morethis right currently falls to your descendantsfor 70 years after your death. Of coursecopyright tends to be used to protect thingsthat have more worth than a scribble;books, plays, paintings and so on areprotected by copyright. For copyright thelaw says that what is important is‘expression of thought’ and not the‘originality of ideas’. In fact copyright legallydoes not require that an expression must be original – just that it is not copied fromanother work.

The Da VinciCopyright

In a recent court case author Dan Brown,writer of the international best-selling book‘The Da Vinci Code`, was taken to court bythe authors of an earlier book `The HolyBlood and the Holy Grail’, because theybelieved that Dan Brown had infringed theircopyright by using similar plot ideas in hisnew book. They lost their case. The Judgedecided that there is no copyright for anidea and the legal protection is for the waythat the idea was expressed. There hadbeen no significant copying between thebooks.

Opportunityknocks for TVGameshows?

A TV gameshow is in effect an idea, arecipe for how to entertain the viewers.These so-called TV formats are big businessworth millions around the world. Can theseideas be protected? With all that money atstake it’s not surprising that there have beenseveral legal cases when one TV station hastried to copy an idea from another and theyhave ended up in court. The first majorcase like this was in New Zealand in 1989.

At the time Hughie Green was the host of a very popular TV talent show calledOpportunity Knocks. A company in NewZealand put on a TV show that the makersof Opportunity Knocks considered to be a copy of theirs, and so infringed theircopyright. The format, their recipe for theshow, was simple, it involved a talent showwith a ‘clapometer’ that measured the noiseof audience applause. The highest scorechose the winning act. They also includedseveral host ‘catchphrases’. Green lost thecase, the courts deciding that there was nocopyright in the format of a talent show.

Big Brother’sBig Day in court

Fast forward to Brazil a few years ago. The reality TV show Big Brother was aninternational hit, and it was being discussedfor a slot on Brazilian TV. The Braziliansdecided not to pay Endemol the creators ofBig Brother to air the show or to buy theFormat Bible, the book that Endemol weretrying to sell them which detailed how tomake the Big Brother show. Instead theycreated a new show called "Casa DosArtistas" (the Artist's House), whichEndemol felt was a "rude copy" of theirshow. The case went to court, Brazilian TVclaiming that a reality show is no more than

Protecting BigBrother’s DaVinci Code

Pub0319 CS4FN4issue4 v4.qxd 21/11/06 15:13 Page 18

Page 19: Special Issue on Computer Science and BioLife · Windows. Arrow and dot diagrams like this are used a lot in computer science to describe how computer systems change over time. Why

19

an idea, and that the idea oflocking folk up and watchingthem wasn’t new either. It hadbeen used in the book "1984" byGeorge Orwell. Even more ironicallythey argued that the lack of scripts,which is what makes reality TV realityTV, meant there was no copyrightprotection to the show.

The cost of agood idea

After much deliberation the Endemol casewas upheld. They had shown that theirformat was more than just the idea, it had a load of identifiable elements to it that hadbeen copied. The expression of the ideasuch as microphones on the housemateslinked 24/7 to the viewers, the music styles,the way housemates had contact with theoutside world, and the tasks, were sufficientthat is was clear the Brazilian show copiedthem directly. The Judges awarded costsand damages of approximately £1.4million.

Software: theplot thickens

So legally ideas, such as those in the HolyBlood and the Holy Grail, cannot beprotected but the way they are expressed indetail (the actual words of the book or theelements of the game show) can stand upin court. So what happens in computersoftware? Computer software is written, notin flowery literary prose or in a Format Bible,but in a computer language. The program,as written, is protected by copyright.However at the core of computer software isthe algorithm, the mathematical recipe ofhow the software does what it is supposedto do. The algorithm is like the ‘plot’, the‘ideas in the book’, the ‘Format’, and thecomputer program is merely the expressionof the ideas like the actual words in thebook. More complex still, algorithms aremathematics, and the law doesn’t do maths very well. Mathematical formulas areconsidered legally as a discovery of theworkings of the universe, rather than aninvention like the paperclip. You can protectan invention, that’s created by a human, butdiscovering the earth goes round the sun, or that a certain mathematical equationdescribes the earth’s orbit is seen as simplyrevealing what’s already there. That’s tworeasons that algorithms can’t be protected.There is a similar controversy over whetherwestern companies can patent products likemedicines based on plants from othercountries where the locals have known the effects for generations: Biopiracy.

Software: a twist in the plot?

The software industry is worth billionsworldwide, and many argue that ifcompanies spend time and money creatingbetter software to solve real problems theyshould be able to protect this investment.Copyright is a major weapon in this war.When you buy a piece of software on a CDyou don’t buy the copyright, you only buy alicense for your own personal use. Softwarepiracy is actually where illegal copies ofsoftware are made, infringing the ownerscopyright and is a major threat to thesoftware industry. A company can spend avast amount of time trying to prevent thecopying of software, but it’s normally only a matter of time till someone cracks thesecurity. Add to this the fact that muchsoftware, to ensure it works effectively,makes use of previously written code in theform of software libraries and you can seewe are in a situation where things get tricky.Software is often a blend of many peoples’intellectual property.

While TV companies can fight in the courts,small software companies have problems inprotecting their work, the expression of theirideas. Some computer scientists believe thatcopyright is the wrong way to go withsoftware, that software should be Open.That is, all the code (at least that theywrite) should be available for others toshare. After all what’s the likelihood thatyour software will still be current and usable70 years after your death?

Stealing Bits

The expression of ideas in the form of digitalbits on a CD or download has taken asmuch intellectual effort, if not more, thanwriting a novel, painting a picture or comingup with the next big TV game show idea.While few of us would ever steal anotherperson’s carefully sanded wooden table orthe Mona Lisa, many still think it OK to copysoftware. The way our society sees thedifference between physical and intellectualproperty, and the tools the law gives us toprotect each, will help define the way thatthe next decades advance. We must hopethat we learn to respect and correctlyhandle computer code in the same way thecourts respect the Da Vinci Code if we wantopportunity to knock for us all in the future.

Email us at cs4fn�dcs.qmul.ac.uk

Feel free to photocopy pages from cs4fn for personal or class use

Pub0319 CS4FN4issue4 v4.qxd 21/11/06 15:13 Page 19

Page 20: Special Issue on Computer Science and BioLife · Windows. Arrow and dot diagrams like this are used a lot in computer science to describe how computer systems change over time. Why

Passionate about computer science?

www.dcs.qmul.ac.uk/cs4fn/

EPSRC support the cs4fn SpaceInvaders project and Sodarace

Back (page) to Life

There are many advances inscience and technology where the innovators involved are way ahead of their time. The technology just wasn’t thereto support their futuristic ideas or the time just wasn’t right. Laterthey were proved to be correct,here we celebrate how, one wayor another, computer science has brought their ideas to life!

Da Vinci coded

One of the most famous examples isLeonardo Da Vinci, the renaissance scientist,engineer and artist. His drawings show thathe thought up many completely novelinventions for his time. He is credited forinventing things like the helicopter, tank,solar power, robots and the calculator. Manyof his inventions have since been made oncemodern materials and tools were available.Computer simulations also showed the leversand strings in his robot knight workedperfectly.

Life Lesson: Even Leonardo’s ideasdidn’t always get off the ground straight away!

Muslim programs

Back in the 9th century in Baghdad thePersian Muslim scholar Abu AbdullahMuhammad ibn Musa al-Khwarizmi wrote a book “On the Calculation with HinduNumerals”. It was responsible for thesubsequent widespread use of the Hindu-arabic number system we use today. He alsodeveloped rules for doing arithmetic usingthis system. The word algorithm, derivedfrom his name, started to be used to refer tosuch rules that could be followed to achievea calculation. Once computers wereeventually invented in the 20th century thiswhole idea of algorithms suddenly becamecrucial as that is really all a computerprogram is: a set of instructions that iffollowed precisely in the given order lead to some task being achieved…but nowfollowed by a computer rather than by a mathematician.

Life Lesson: Always have a planand stick to it!

Grumpy old men can

make a difference

In the 1800s “computers” were teams ofpeople employed to calculate tables ofnumbers used for example when navigating.Charles Babbage changed that inventing thefirst programmable computer, called “TheDifference Engine” to take over and soeradicate human error from the table.Unfortunately he struggled to get the fundingfor his ideas not least because of hisconfrontational personality and he never builtit. The world had to wait a century longer fora functioning computer. In 1991 a versionwas finally built directly from his plans – and worked perfectly.

Life Lesson: Success takespersonality too!

The pen, the paperand the poet’s

daughter

Babbage’s computers needed programs andfor that he needed a woman: Ada Lovelace,daughter of poet Byron and an accomplishedmathematician. Even though the DifferenceEngine did not exist she wrote programs forit, and even tested them on paper to makesure they did work even though she couldn’trun them on the machine. The sametechnique is still used today by programmersto help get rid of bugs in their code at anearly stage.

Life Lesson: You don’t need afamous poet dad or a computer to teachyourself to program.

Small but perfectly

formed

In 1960, Richard Feynman, Nobel prizewinning physicist, arguing that we were along way off the fundamental limit of howsmall machines could get (after all biological“machines” like cells manage at molecularscales) challenged scientists and engineersto make nano-machines. This helped triggera whole new area of computer science and isstarting to achieve results: with nano-enginesand messages written in atoms that can beread only with atomic force microscopes.Computers the size of specks of dust are not far away. They will be everywhere, but how will we use and program them?

Life Lesson: There is plenty of room for computers at the bottom, top, sides, edges….

Computer say ‘No!’

In 1990, entrepreneur Hugh Loebner offereda prize for the first piece of software to passa variation of a test first suggested by AlanTuring in 1950 as a way to see if artificial"intelligence" had been achieved. To passthe test and so win the prize the winningchatbot program must convince a panel of experts that they are conversing with a human not a chatbot. So far ArtificialIntelligence eludes us and the prize isunclaimed: recently Loebner claimed “At this rate, I will be dead before the Turing test is passed”.

Life Lesson: Chatting is far harderthan you think, just ask any computer.

Microsoft, ARM and Intel provide support for printing.

See the webzine’s Magazine+

www.dcs.qmul.ac.uk/cs4fn

for extended versions of this

and other articles

Back (page) to Life

cs4fn is produced by Paul Curzon and Peter McOwan of the Department of Computer Science, Queen Mary, University of London. Support provided by Jane Reid, Gill Carter, Ella Rice and Sue White.

This issue also includes a contribution by members of Bristol University - see page 4

Pub0319 CS4FN4issue4 v4.qxd 21/11/06 15:13 Page 20