Monograph: Computing Omnipresence (published … The Open Source Software vs. Proprietary Software...

* This monograph will be also published in Spanish (full version printed; summary, abstracts, and somearticles online) by Novática, journal of the Spanish CEPIS society ATI (Asociación de Técnicos deInformática) at <http://www.ati.es/novatica/>, and in Italian (online edition only, containing summary,abstracts, and some articles) by the Italian CEPIS society ALSI (Associazione nazionale Laureati in Scienzedell’informazione e Informatica) and the Italian IT portal Tecnoteca at <http://www.tecnoteca.it>.

Vol. VI, issue No. 5, October 2005

UPGRADE is the European Journal for theInformatics Professional, published bimonthly

at <http://www.upgrade-cepis.org/>

UPGRADE is the anchor point for UPENET (UPGRADE EuropeanNETwork), the network of CEPIS member societies’ publications, that currentlyincludes the following ones:• Mondo Digitale, digital journal from the Italian CEPIS society AICA• Novática, journal from the Spanish CEPIS society ATI• OCG Journal, journal from the Austrian CEPIS society OCG• Pliroforiki, journal from the Cyprus CEPIS society CCS• Pro Dialog, journal from the Polish CEPIS society PTI-PIPS

PublisherUPGRADE is published on behalf of CEPIS (Council of European ProfessionalInformatics Societies, <http://www.cepis.org/>) by Novática<http://www.ati.es/novatica/>, journal of the Spanish CEPIS society ATI (Asociaciónde Técnicos de Informática, <http://www.ati.es/>)

UPGRADE monographs are also published in Spanish (full version printed; summary,abstracts and some articles online) by Novática, and in Italian (summary, abstracts andsome articles online) by the Italian CEPIS society ALSI (Associazione nazionale Laureatiin Scienze dell’informazione e Informatica, <http://www.alsi.it>) and the Italian ITportal Tecnoteca <http://www.tecnoteca.it/>

UPGRADE was created in October 2000 by CEPIS and was first published by Nováticaand INFORMATIK/INFORMATIQUE, bimonthly journal of SVI/FSI (Swiss Federation ofProfessional Informatics Societies, <http://www.svifsi.ch/>)

Editorial TeamChief Editor: Rafael Fernández Calvo, Spain, <[email protected]>Associate Editors:François Louis Nicolet, Switzerland, <[email protected]>Roberto Carniel, Italy, <[email protected]>Zakaria Maamar, Arab Emirates, <Zakaria. Maamar@ zu.ac.ae>Soraya Kouadri Mostéfaoui, Switzerland,<[email protected]>

Editorial BoardProf. Wolffried Stucky, Former President of CEPISProf. Nello Scarabottolo, CEPIS Vice PresidentFernando Piera Gómez andRafael Fernández Calvo, ATI (Spain)François Louis Nicolet, SI (Switzerland)Roberto Carniel, ALSI – Tecnoteca (Italy)

UPENET Advisory BoardFranco Filippazzi (Mondo Digitale, Italy)Rafael Fernández Calvo (Novática, Spain)Veith Risak (OCG Journal, Austria)Panicos Masouras (Pliroforiki, Cyprus)Andrzej Marciniak (Pro Dialog, Poland)

English Editors: Mike Andersson, Richard Butchart, David Cash, Arthur Cook,Tracey Darch, Laura Davies, Nick Dunn, Rodney Fennemore, Hilary Green, RogerHarris, Michael Hird, Jim Holder, Alasdair MacLeod, Pat Moody, Adam David Moss,Phil Parkin, Brian Robson

Cover page designed by Antonio Crespo Foix, © ATI 2005Layout Design: François Louis NicoletComposition: Jorge Llácer-Gil de Ramales

Editorial correspondence: Rafael Fernández Calvo <[email protected]>Advertising correspondence: <[email protected]>

UPGRADE Newslist available at<http://www.upgrade-cepis.org/pages/editinfo.html#newslist>

Copyright© Novática 2005 (for monograph and cover page)© CEPIS 2005 (for editorial, and sections MOSAIC and UPENET)All rights reserved. Abstracting is permitted with credit to the source. For copying,reprint, or republication permission, contact the Editorial Team

The opinions expressed by the authors are their exclusive responsibility

ISSN 1684-5285

Monograph of next issue (December 2005)The Semantic Web

(The full schedule of UPGRADEis available at our website)

2 EditorialIn The Fifth Anniversary of UPGRADE — Geoff McMullen (Presi-dent of CEPIS)

3 PresentationAn Outlook for Ubiquitous Computing: Computing Power Anywhereand in Any Device — José-Antonio Gutiérrez de Mesa, DanielRodríguez-García, and Miltiadis D. Lytras

8 The Critical Mass Problem of Mobile Ad-hoc Networks — Jörg Roth14 Device and Context Influence on Wireless Infotainment Access: A

Real World Story — Tacha Serif and George Ghinea20 The Problems of Printing in Server-Based Computing Environments

— Luis Bengochea-Martínez26 Using FOAF to Support Community Building — Brian Kelly and

Leigh Dodds29 Towards Ubiquitous Computing with Quality of Web Service Sup-

port — Yannis Makripoulias, Christos Makris, Yiannis Panagis,

Evangelos Sakkopoulos, Poulia Adamopoulou, Maria Pontikaki, and

Athanasios Tsakalidis35 The Open Source Software vs. Proprietary Software Debate and Its

Impact on Technological Innovation — Ricardo-José Rejas-Muslera,Juan-José Cuadrado-Gallego, Javier Dolado-Cosín, and DanielRodríguez- García

39 Localization in Ubiquitous Computing Using Acoustic Sensor Net-work — Carlos-Manuel De Marziani, Jesús Ureña-Ureña, ÁlvaroHernández-Alonso, Manuel Mazo-Quintas, Ana Jiménez-Martín,Juan-Jesús García-Domínguez, José-Manuel Villadangos-Carrizo,and Fernando-Javier Álvarez-Franco

46 Portable System For Patient Monitoring With Wireless Technologies— José-Manuel Rodríguez-Ascariz, Luciano Boquete-Vázquez,Ignacio Bravo-Muñoz, and Pedro Martín-Sánchez

53 SIGLAS: A Practical Case of The Application of Ubiquitous Com-puting in Warehouse Management — José-Julio González-Pozo andManuel Ortega-Cantero

58 Ubiquitous E-Management of Indicators — Eladio Domínguez-Murillo, Ángel De Miguel-Artal, Beatriz Pérez-Valle, ÁureaRodríguez-Villanueva, and María-Antonia Zapata-Abad

61 From Pro Dialog (PIPS, Poland)Computer Algebra SystemsProblems with A Heart-Shaped Curve — Leszek Jankowski and AdamMarlewski

68 From Mondo Digitale (AICA, Italy)Web EngineeringA Quality Model for Websites — Roberto Polillo

UPENET (UPGRADE European NETwork)

Monograph: Computing Omnipresence (published jointly with Novática*)Guest Editors: José-Antonio Gutiérrez de Mesa, Daniel Rodríguez-García, and

Miltiadis D. Lytras

UPGRADE Vol. VI, No. 5, October 2005 61© CEPIS

UPENET

Computer Algebra Systems

Keywords: computer algebra Systems, Lambert func-tion

1 Introduction: NOS, SOS, and Errors in SOSesIn the beginning computers handled nothing but num-

bers, also if represented by letter symbols, later called‘names’ or ‘identifiers’. And, although very soon they alsoprocessed text (usually called ‘strings’), by such operationas the deletion of a character and the concatenation, theywere designed and used to process numbers. The systems(a.o. BASIC, FORTRAN, Algol, Pascal, C++) intentionallymanipulating numbers are called NOSes, i.e. numericallyoriented systems.

First trials to analytically handle derivatives and inte-grals on computers were undertaken in 1953 (for more de-tails see e.g. [3] and [11]). First programs intentionally de-signed to be a direct mathematical aid to scientists in theirdaily work were released in 1968: MATH-LAB-68 andREDUCE. They both factorized polynomials, calculatedindefinite integrals, and solved some differential equations.These two systems (called SOSes, i.e. symbolically orientedsystems, or CASes, i.e. computer algebra systems) were soonfollowed by such products as MATHSCRIBE (1969, the firstWYSIWYG symbolic algebra programs, also showinggraphical animation), MACSYMA (1969), SCRATCHPAD(1974) turned into Axiom, MuMath (1979, the first sym-bolic algebra program to run on microcomputers managedby CP/M, an operating system for 8-bit computers, the pred-ecessor to DOS) and its successor DERIVE (1988), alsoimplemented in hand-held computers TI-89, Maple (1983),and Mathematica (1988). There are many other SOSes, inparticular designed to deal within specialised areas (e.g.

Cayley, Fermat, GAP, PARI/GP; see [10]), but it was Ma-ple, Mathematica and DERIVE (now from Waterloo MapleInc. a.k.a. Maplesoft, Wolfram Research, Texas Instruments,respectively) that gained the biggest popularity in educationand university research. There are also popular hybrids suchas MathCAD and MATLAB (from Mathsoft andMathWorks, resp.); they are basically NOSes, but they canmanipulate symbolical transformations if equipped withspecial packages.

Problems with A Heart-Shaped Curve

Leszek Jankowski and Adam Marlewski

© Pro Dialog 2005This paper was first published, in English, by Pro Dialog (issue no. 19, 2005, pp. 135-146). Pro Dialog, <http://www.pti.poznan.pl/prodial/

prodialEn.html>, a founding member of UPENET, is a journal copublished, in Polish or English, by the Polish CEPIS society PTI-PIPS (PolskieTowarzystwo Informatyczne – Polish Information Processing Society) and the Poznan University of Technology, Institute of Computing Science.

We examine the work of two computer algebra systems on an ordinary differential equation of the first order. Both systems,DERIVE 5 and Maple 8, yield implicit solutions, and both have problems plotting graphs of integral curves (which may beclassified as heart-shaped). These problems are even more serious when, in Maple, we deal with the implicit solution,where the Lambert function is involved.

Leszek Jankowski obtained, in 1983, a PhD in Mathematicsfrom the Adam Mickiewicz University in Poznan, Poland, wherehe presented the dissertation on automorphisms of semigroupsand their application in modular spaces. For several years heworked as Solidarity treasurer in Gdansk, Poland. Since 2002he has worked again for Poznan University of Technology. Hisrecent research work is focused on topology. His personal pageis at <http://www.math.put.poznan.pl/people/ljankows.html>.

Adam Marlewski graduated in numerical methods at the AdamMickiewicz University in Poznan, Poland. In 1981 he got a PhDdegree for a thesis on Bernstein polynomials. Since 1973 he worksin the Poznan University of Technology, and in 1992-94 he lecturedin the Universidade da Beira Interior in Covilha, Portugal. Hehas published some books, authored and co-authored papers inapproximation theory, computer algebra systems, numericalmethods, electrical machines and simulation in remote sensingand civil engineering. He is a reviewer in Mathematical Reviews.He heads a site project on applications of symbolic calculations.His personal page is at <http://www.math.put.poznan.pl/~amarlew>. <[email protected] >

This section includes articles published by the journals that make part of UPENET.For further information see <http://www.upgrade-cepis.org/pages/upenet.html>.

62 UPGRADE Vol. VI, No. 5, October 2005 © CEPIS

UPENET

Obviously, the most important feature of CASes is thatthanks to them researchers and students may be freed fromlaborious, time-consuming transformations and calculationswhich are performed in an exact mode, i.e. without the round-ing which are essentially involved in NOSes. Although in-tensively developed, CASes are still (and probably they willalways be) only assistants to mathematical investigations.This is expressively announced in the full name of DERIVE:DERIVE. A mathematical assistant. It is DERIVE, version5, which is a CAS, we make use of. By the way, in 2001Derive 5 was voted Best Budget Buy of the mathematicsection by the English magazine Computer Shopper (the di-dactic use of DERIVE and other CASes is treated in numer-ous publications, e.g. [1] – [3], [11], [13] – [15], [17] – [18],[20], [21], [23], [24]).

And although today’s CASes do much more than theirfirst predecessors, there are still problems which CASes finddifficult to deal with and bugs that users find in them. Thishappened recently to a version of Maple, where the expres-sion 2001!/2000! simplified to 1. DERIVE, which returns acorrect result, provides an obviously wrong answer –3/2when we ask for the value of

One easily notices that this error is of a different typefrom that yielding 2001!/2000! to 1, The former is simply atechnical mistake, and the latter is made by an unjustifiedapplication of Leibniz’s formula for definite integrals. Inour paper we discuss difficulties DERIVE and Maple havewhen an ODE (ordinary differential equation) is to be solvedand its solution is to be represented graphically.

2 Heart-Shape CurvesLimaçons of Pascal are covered by the polar equation

and were investigated by Albrecht Dürer (1471–1528) andÉtienne Pascal (1588–1651). A special case for b = 2a is acurve traced by a point on the circumference of a circlerolling round the circumference of a circle of an equal ra-dius. In 1741 this epicycloid was named a cardioid by JohannCastillon. He derived this name from Greek word kardi(heart), this way he wanted to say that this curve is a heart-shaped. The cardioid is not only a mechanical curve whichwould make it possible to construct a device which moves atappet sinusoidally. The cardioid is also the envelope of lightbeams emitted from a point laying on a circle and reflectedfrom this circle. Its good approximation can be seen on thesurface of a mug filled with coffee.

The cardioid, in spite of its name, resembles a heart in amuch lesser degree than other lines. For example, we getmore natural heart-shaped curves when working with someBonne projections (see [4]). Other examples of heart-shapedlines are

and

with θ running from to , first two onesproposed in 2003 by Dascanio [9], last two ones listed byK. Eisemann of San Diego University [5], or projections onOxy plane of the surface

worked out 11 years ago by Gabriel Taubin of IBM [22],and

Fig. 1. Graphs of equalities x ln y – x2 – y2 = c forc = –2,–5, –8

Fig. 2. Graphs of equalities x = (ln y – v)/2 andx = (ln y+ v)/2, where v = and u = ln2u – (v3 – 8)


UPENET

modeled in 1997 by T. P. Nordstrand of Bergen Univer-sity [19] (see also J. P. Kuska of Leipzig Universitaet [12],[16]). In the next section we find an other heart-shaped curve.

3 Solving an ODE in DERIVEWe look for the solution y = y(x) of the differential equa-

tion

(1)

Our equation is an ODE1, i.e. ordinary differential equa-tion of the first order, which has the form p + qy’ = 0, wherep and q are expressions depending eventually on x and y.That’s why in DERIVE we can evoke the procedureDSOLVE1_GEN. Hence we edit the expression

DSOLVE1_GEN(LN(y)–2x,x/y – 2y, x, y, c). (2)

It simplifies to x·LN(y) – x2 – y2 = c, (3)where c stands for an arbitrary constant (named an inte-

gration constant).For various values of the constant c we may let DERIVE

plot the solution, i.e. we may get the graphs composed ofpoints (x, y) satisfying the final relation (3). In Fig. 1 we seethe graphs issued by DERIVE for c = – 2, – 5, and – 8.

It is easy to see that these graphs are far from perfection.The relation (3) is essentially nonlinear with respect to

y. But we can easily express x in terms of y. We can let it bedone by DERIVE as follows: we simplify the call

SOLVE(x·LN(y) – x2 – y2 = c, x). (4)

We get the answer saying

x = (LN(y) – v)/2 OR x = (LN(y) + v)/2 (5)

where v = , u = ln y 2 – 4 (y2 + c).

In Fig. 2 the resulting equalities for c = – 8 are plotted.In contrast to Fig. 1, graphs approach the horizontal axisOx, and there is a gap between left and right branches at theapproximity of the point (x, y) = (0.6, 2.85). Should theycompose a continuous curve?

Recalling the formula , and DERIVEcounterpart of this relation, INT(1/x,x) = LN(x), onecould suspect that the absolute value sign can be omitted inthe yielded answer. In fact it is not forgotten, DERIVE simplydoes not displays an integral constant. If we assume it isequal to, let’s say, –iπ + c with arbitrary positive c, we haveln| x |+ c). It appears that we are right, so we can modify theanswer (3) to the form

x·LN(|y|) – x2 c y2 = c. (6)

The image of this relation with c = – 8 is shown in Fig.3. It reveals a "strange" plot close to the horizontal axis Ox.The equation (6) says that this axis has to be the symmetryaxis (replacing !y for y changes nothing), so we know thatthe graph is only a sketchy one (and its quality may be influ-enced by the screen resolution).

We have produced three graphs and not one of them canbe surely accepted. Nevertheless, these graphs show thatintegral curves may be heart-shaped, that they resemble astylized heart spread along the horizontal axis. We wouldlike to know if these graphs are really heart-shaped or not.As shown above, we have doubts about the graphs of thesolution (5) to the problem (1) when y approaches 0 andwhen y is close to 2.85 (and x ≈ 0.6, as it is revealed in Fig.2). The first question can easily be handled by the limitingprocedure which is built-in in DERIVE. We edit the expres-sionsLIM((LN(y) – SQRT(LN(y)^2 – 4A(y^2 +c)))/2, y,0),LIM((LN(y) + SQRT(LN(y)^2 – 4A(y^2 +c)))/2, y, 0),

Fig. 3. Graph of the equation (6) with c = – 8

Fig. 4. Graphs of expressions y= ln y, y= ln2x, and y= 4(x2 –8), and points E= (e, 1) and M= (q, 4(q2 – 8)), where the lasttwo graphs cross


UPENET

Fig. 5. Screen displayed by Maple: solution to the ODE1 considered and the attempts to check it(the direct one and via the odetest function)


UPENET

the first call to get the limit for the left branch, and thenext call to know the right branch of our solution (5).

Unfortunately, simplifications of both these calls fail asDERIVE does not remove the LIM operator. It does removeit if we command the right-side limiting, or if we replace yby | y |. Then the calls

LIM((LN( y ) – SQRT(LN( y )^2 – 4A(y^2 +c)))/2, y,0,1), LIM((LN(|y|) – SQRT(LN(|y|)^2 – 4A(y^2 +c)))/

2, y, 0)

simplify to .

If we dare to say that our integral curve is heart-shaped,we have to be aware that this heart has "a tail", that thisheart is chained, as if tied by a string spreading along thenegative semiaxis Ox.

We are successful if we limit from the right or we limitwith y replaced by | y |. Unfortunately, an analogical processdoes not work when we deal with the limit concerning theright branch, DERIVE returns an expression which differsfrom the given one only by cosmetic changes (and with |y|the outcome is even less clear).

What DERIVE cannot do, we can do. A simple comple-tion of the sum by the difference to have the difference ofsquares gives

If y → 0+, the nominator approaches the value 4c andthe denominator . Hence

Fig. 6. Direction field of the equation (1) and the graph of one of its solutions plotted in DERIVE

This proves that the right branch approaches 0 when ydecreases to 0 (and it runs to 0 by positive values if c < 0).So our "heart" has a deep indentation, which it reaches theorigin O.

At last, we discuss do both branches touch or are theydisjoint curves. By their definition it has to be

y > 0 and ln2 y – 4(y2 + c) ≥ 0.The last inequality cannot be solved in an exact way,

because the equationln2 y – 4(y2 + c) = 0

is transcendent. It determines a number, let’s call it q,such that both expressions

and

are defined for y ∈ ( 0, q]. Putting y = q in both theseexpression we get the same value lnq. This proves that leftand right branches meet. Fig. 4 illustrates the situationfor c = –8. The value of q, y = 2.87736, is returned by thesimplification of the call

NSOLVE(LN(y)^2–4*(y^2–8),y).

4 Explicit Form of the SolutionThe Lambert function, also called the omega function,

and denoted here by the letter W, is defined to be themultivalued inverse of the function z → z exp z. This meansthat W satisfies the equation

W(z) exp(W(z)) = z (7)

For z <0 the W function has only two real-valuedbranches. The branch with values greater than –1 is calledthe principal one. This function was named after JohannHeinrich Lambert (1728 –1777) who was the first to pro-

Propietario

Placed Image


UPENET

Fig. 7. Graph produced in Maple

vide, in 1761, a rigorous proof that π is irrational1 . In someof his works he dealt with the expressions satisfying therelation (7), but a real interest concerning it exploded onlysome years ago, and recently many applications of this func-tion have been shown, see e.g. [6], [7], and [8].

By means of the W function we can provide an explicitform of the solution of our equation (1). For this purpose letus take a new variable t = ln y. Then the solution (3) takesthe form

xt – x2 – exp(2t) = c, i.e. 1 = (xt – x2 – c) exp(–2t).

That is why vu exp u, where

Thus, by the definition of the Lambert W function,u = W (v)i.e.

where

Finally,

and so, if

then the explicit form of the solution to (1) is

In DERIVE the Lambert function is not implemented,whereas in Maple it is identified by LambertW.

5 Looking for a Solution in MapleAs we have already said, the Lambert function is imple-

mented in Maple 8, and it is identified by LambertW. Thesimplification of the call

dsolve(iod)

where iod stores the considered equation, i.e.

iod := ln(y(x))–2*x+(x/y(x)–2*y(x))*diff(y(x),x),

yields the expression of the form (8), the only differencebeing that instead of c, the integration constant is –_C1. Onecan see it in Fig. 5, where the attempts to check whether theoutput expression really solves the equation (1) are alsoreproduced. The direct checking, consisting in the substitu-tion of the obtained result into the equation (1), is notsatisfactory as we get a lengthy and illegible expression, andwe can even suspect that (8) does not solve (1). Theconfirmation that (8) solves (1) is provided by the call ofthe built-in Maple function called odetest, intentionallydesigned to be applied in such situations.

As we can see in Fig.7, the graph, produced in Maple, ofthe solution satisfying the condition y(x = 1) = 1, is incom-plete: it does not pass through the point (x, y) = (1, 1) nordoes it pass in its neighborhood. The graph of the relation(3), providing the implicit solution to (1), however passesthrough this point. We see it in Fig. 6, produced in DERIVE,where the direction field of (1) is shown. The result is sur-prising: despite the enrichment of the system (Maple withthe function LambertW) we get a result which confuses us.

This conclusion once again confirms that CASes are onlymathematical assistants. It stays valid also when, in DERIVE,instead of the call

1 121 years later C. L. F. von Lindemann established that π istranscendental.


UPENET

DSOLVE1(LN(y) – 2·x, x/y – 2·y, x, y, 1, 1)

we simplfy the expression

DSOLVE1( –(LN(y) – 2·x)/(x/y – 2·y),1, x, y, 1, 1).

The first call refers to the equation (1), the second one tothe equation

It is different from (1) only on the curve x = 2y2 which,obviously, is not a solution to the Cauchy problem (1) or (2)with y(x = 1) = 1. As we know, DERIVE provides the solution(3) in the first case. The simplification in the second case fails,as the program displays the information: inapplicable.

References[1] Adamczak, M., Using New Technologies in Maths

Education at University Level [in Polish], Ph. D. The-sis, Adam Mickiewicz University, Poznañ 2002.

[2] Ben-Israel, A., Gilbert, R. P., Koepf, W., Mathematikmit Derive, Vieweg, Braunschweig/ Wiesbaden 1993.

[3] Bogolubskaja, A., Computer Algebra Systems [in Rus-sian], 2001, http://ca-d.jinr.ru/literature/tutorial/cas.htm.

[4] (Bonne projection) http://mathworld.wolfram.com/BonneProjection.html

[5] (Cardioid) http://www-gap.dcs.st-and.ac.uk/~history/Curves/Cardioid.html

[6] Corless, R. M., Gonnet, G. H., Hare D. E. G., Jeffrey,D. J, Knuth, D. E., On the Lambert W Function, Adv.Comput. Math. 5 (1996), 329–359.

[7] Corless, R. M., Jeffrey, D. J., Knuth, D. E., A Sequenceof Series for the Lambert W Function, in: Proc. ISSAC’97, Maui, Hawaii (Ed. W. W. Küchlin), ACM, NewYork 1997, 197–204.

[8] Cranmer, S. R., New Views of the Solar Wind with theLambert W Function, Am. J. Phys. 72 (11) 2004, 1397–1403.

[9] http://mathworld.wolfram.com/HeartCurve.html.[10] Workshop on Open Source Computer Algebra, Report

by Daniel Duparc et al., 2002, http://lapcs.univ-lyon1.fr/~nthiery/CalculFormelLibre/report/.


UPENET

Keywords: Quality Model, SoftwareQuality, Web Assessment, Web Evalu-ation, Web Quality.

1 IntroductionEven if web technologies allow

users with little experience in Compu-ter Science to quickly create elemen-tary websites, designing professionalwebsites is a very complex process. Infact, there are many different aspectsto keep in mind. The website must bein line with the communication style ofthe organization it represents, it mustpresent the characteristics and the prod-ucts or services in an accurate way.More and more often it is necessary forthe visitor to be able to do complexoperations, such as purchases, reserva-tions, access to databases, which re-quire a close integration with companyinformation systems. Also during thefunctioning of the website, it is impor-tant to manage complex problems. Forexample, to make sure that the informa-tive contents are always and timelyupdated, that the technological infra-structures used guarantee acceptableservice and run correctly and continu-ously, or that information requests fromvisitors are answered.

All the above require the use ofpersonnel with specific professionalskills which up to a few years ago didnot even exist. Furthermore, the rangeof competences needed to create a qual-ity website is large. The contributionof different people is thereforenecessary; experts in marketing, com-

munication, software, and above allexperts in the specific business of thecompany that owns the website. Theoverall quality of a website depends onthe harmonious cooperation of all theprofessionals involved: the results ofthe work of each much integrate cor-rectly with the results of the work ofall the others. No contribution can beignored: because, like in a chain, glo-bal quality is equal to the quality of theweakest link. These needs are normallyunderestimated today: in the implemen-tation of projects, or the improvementof a website, the focus is only on someof the aspects involved, neglectingothers, equally important to the eyes ofthe final user. It is precisely the finaluser who ultimately determines thesuccess or failure of a website.

To evaluate the quality of an exist-ing website, or to correctly plan a newone, it can therefore be very useful touse a simple quality model which re-minds us of the principal factors in-volved and that allows us to keep themunder control. This article syntheticallydescribes a quality model explicitlydesigned for websites, which was usedsuccessfully on websites of every kindand complexity, both for making a rapidcheck-up and also for in-dethassessments. The model, which is de-scribed in depth in a recent book bythe author of this article [5], takes intoaccount, above all, the quality per-ceived by the users of the website; inother words, the so-called externalquality and quality in use, accordingto the ISO (International Organization

for Standardization) [2] terminology,leaving aside the aspects of the inter-nal structure of the software used in thewebsite (internal quality).

A quality model can be defined inmany ways. However, it is very con-venient - as we will see in the follow-ing - to start from the analysis of theactivities involved in the design, thecreation and management of a website,and of the various types of actors in-volved. Therefore, in Section 2, thelogical phases that occur in the crea-tion and operation of a website arebriefly described. Starting from this, inSection 3, the quality model is intro-duced, which will be further analyzedin Section 4. In Section 5 we describehow the model can be used in the evalu-ation of websites. Following is a briefconclusion (Section 6), and an essentialbibliography of the issues discussed.

2 How A Website Is BuiltWeb engineering is a recent sub-

Roberto Polillo teaches Man-MachineInteraction at the Università degli StudiMilano Bicocca, Italy. For over thirtyyears he has worked in the ComputerScience field, in particular software, asa professor, entrepreneur and manager.He is one of the founders of Enoteam, awell-known Italian ICT servicescompany, for which he was CEO duringthe boom of Internet. Apogeo recentlypublished one of his books which des-cribes in detail the methodology forevaluating the quality of a website, whichhas been summarized in this article.<[email protected]>.

Web Engineering

A Quality Model for Websites

Roberto Polillo

© Mondo Digitale, 2005This paper was first published, in its original Italian version, under the title “Un modello di qualità per i siti web”, by Mondo Digitale (issue no. 2,Giugno 2005, pp. 32-44, available at <http://www.mondodigitale.net/>). Mondo Digitale, a founding member of UPENET, is the digital journal ofthe CEPIS Italian society AICA (Associazione Italiana per l’Informatica ed il Calcolo Automatico, <http://www.aicanet.it/>.)

The quality of a website is a product of the activity of many actors: web designers, graphic designers, software developers,content writers, webmasters, and web managers. The result of the work of each specialist must integrate correctly with the workof the others. In order to evaluate the quality of a website, a “quality model” is very useful, since it reminds us of the principalfactors involved in achieving it. This article describes a simple quality model for websites, which has been used widely.


UPENET

ject and is not yet consolidated, and inpractice there are varied approachesand methodologies, normally not verystructured, for planning and creatingwebsites. A universally adopted modelof processes does not exist. Also be-cause, for example, the creation of alarge portal is very different (it can in-volve a project group with tens of peo-ple) from that of an institutional websiteof a small company (which could becreated and managed by a singleperson, who can take care of all theaspects).

Without pretending to be dealingwith a universal model, we briefly de-scribe the general phases which needto be followed from the conception,implementation and management of awebsite in order to identify the type ofactors involved in each phase and thecontributions that each member givesto the overall quality of the website.

This schema deals mainly withwebsites having a certain level of com-plexity, which therefore require theinvolvement of many people: e-com-merce websites, internet banking, pub-lic administration websites, and othersimilar types of websites. In the caseof simpler websites, the model is stillvalid if the necessary simplificationsare made: the roles which are describedwill not necessarily be taken care of bydifferent people. In a simpler website,is not unusual to have only one personin charge of everything.

The various phases that appear inFigure 1 show the principal activitieswhich must occur for creating and run-ning a website, as well as the logical

relationships between them, and theprofessional skills required for eachactivity.

For example, the web design phasereceives its input from the definition ofthe strategy; the visual design requiresthat the structure of the website hasalready been defined in the web designphase; the management of contentsrequires that a software container bealready available, and so on.

Using Figure 1 as a reference, theprocess begins at the planning of thegeneral strategy and requirements,which consists of defining the objec-tives associated to the website, identi-fying the target audience, the require-ments and the constraints that must bemet, the services that the website mustprovide to users, the context of use andpossibly the technological architectureto be used. Typically, in this phase, ananalysis of competitors’ websites sin-gles out the best practice in the busi-ness area and allows for the definitionof an appropriate differentiation policy.Usually this preliminary phase is han-dled by the company’s top managementwith the help of consultants specializedin this field. The output of this phase isa document specifying the generalrequirements, often used by thirdparties in order to build technical andeconomic proposals for the executionof the following phases.

The next phase shown in the figureis web design. In this case the generalinformation architecture of the websiteis designed, and, most of all, the navi-gational structure, defining the interac-tion modalities between the website and

the users, and sketching the generalcharacteristics of the visual appearanceof the pages. Web designers areprotagonists of this phase, architects ofthe web who often work in organiza-tions called web agencies.

Next comes visual design, in whichthe graphics of the website are detailed,on the base of the planning made in theprevious phase. Sometimes variousprototypes are built, in order to be ableto choose the best solution. This phaseis carried out by the visual designer,designers with a strong graphics andvisual communications backgroundwho work closely with the webdesigner.

The software development of thewebsite is the next phase, where stand-ard software products are used often.This is the activity which is closest to"classic" information systems develop-ment and, in the case of less complexwebsites, can be carried out by the sameweb agency that intervened in theprevious phases. Instead, for morecomplex websites these activities arenormally led by companies specializedin system integration. They assemblethe standard software products anddevelop specific software componentswhen necessary, and normally take overthe project, in cooperation with the webagency (if one has been also involvedin the project). The softwaredevelopment activities prepare, so tosay, an "empty" website (without infor-mation contents). Content editors workon the contents aspect in the nextlogical phase, contents management.Basically it involves editors who are

Figure 1: Activities and Professional Skills Involved in The Creation and Management of A Website.


UPENET

Figure 2: The Construction of A Website Is An Iterative Process.

expert in the business domain thewebsite belongs to, and carry out theirwork not only in this phase but alsowhen the website is on line.

When the website is online, it isnecessary to manage and supervise itsoperation. This task is performed bywebmasters, who have the responsibil-ity of making numerous and frequenttechnical interventions in order to keepthe website running and updated, pre-venting interruptions in service. In thecase of websites which offer complexservices, it is also necessary to managethe information systems the websitecooperates with for the execution of theservice. For example, in an e-commercewebsite, the application software forthe management of warehouse,accounting, and any functions relatedto the sale and shipping of goods. In

these cases, the webmaster shouldreceive support from specialists in theapplications involved.

Next in Figure 1 diagram comes themanagement of the server whichhosts the website. This activity can bedone internally by the company thatowns the website of by an external webfarm, if the decision of delegating viaan outsourcing contract has been made.There are firms who manage largequantities of web servers for variousclients. They have the appropriate in-frastructure for this task, allowing sig-nificant savings to their customers infields like management and mainte-nance of computers operations, secu-rity and continuity of service.

The last actor present in the diagramis the Internet Access Provider (IAP),which is the organization which

supplies Internet connectivity to theweb farm: an IAP is basically a tel-ecommunications operator who pro-vides access to the Internet and theservices associated to this access.

We have identified all the activitiesrelative to the planning and manage-ment of a website, their interrelationsand the key actors. These must worktogether, each one playing its own role,in order to guarantee proper operationof the website. Their activities are verydifferent from each other and are ledby professionals from different areas,with very different work methods, lan-guages, professional cultures and val-ues.

The logical sequence of the differ-ent activities does not necessarily al-ways coincide with the time sequence;in fact, the process of creation of awebsite, as in every complex softwaresystem, rarely follows a linear path.Often changes to decisions alreadytaken occur, which produces modifica-tions to the work already done, in aniterative process which converges, soto speak, through subsequent approxi-mations (Figure 2). In any case, thebasic structure of the website isplanned, created, presented to the cus-tomer and tested with real or simulatedusers, customers’ and users informationis collected, and the project is refinedwith this information, improving theunsatisfactory parts. When the websitebecomes operational, its evolution isnot finished yet, but continues for therest of its life span. A website which is"alive" is improved continuously andis adapted to new needs, not onlyupdating and enriching it with new

Table 1: A Tentative Quality Model for Hotels.

Characteristic Sub-characteristics 1 Rooms Comfort, size, elegance, cleanliness 2 Public areas Comfort, size, elegance, cleanliness 3 Service Reception, room service, staff courtesy 4 Restaurants Food quality, environment, service, food variety 5 Swimming Pool Comfort, size, environment, equipment, cleanliness 6 Beach Distance from hotel, equipment, landscape 7 Surroundings Possibility of excursions, distance from city 8 Sports facilities Tennis, golf, fitness, other 9 Children Nursery, playing areas


UPENET

contents, but also making modificationsand improving the software "container"and the hardware infrastructure whichhosts it.

3 A Quality Model for WebsitesHaving described the different

phases and activities which must takeplace in the creation and managementof a website via a global schema, wewill now introduce the basic elementsfor the quality model which is the sub-ject of this article.

It is useful, above all, to define whatis intended for quality model. When thequality of something needs to be evalu-ated, it is necessary to decide whichcriteria have to be used. For example,in order to evaluate the quality of ahotel, a given level of importance mustbe assigned to the elements which areconsidered important, for example:rooms, public areas, service, restau-rants, swimming pool, beach, surround-ings, sports facilities, children’s’ areas,and so on. The same technique mustbe used in order to define the criteriafor evaluating every single element. Forexample, the quality of the rooms couldbe evaluated on the basis of comfort,size, elegance and cleanliness. A set ofcriteria is shown in Table 1, which is atentative quality model for hotels. Thenine basic elements chosen are calledcharacteristics, the others, sub-characteristics.

It is evident that the characteristicsand sub-characteristics of any qualitymodel can be defined in many differ-ent ways, according to the point of viewthat is chosen. In the case of hotels, forexample, we may adopt the point of viewof the company that providesmaintenance services, or the tour opera-tor who sells the service to the public, orthe client. Depending on the point ofview, the model will be different.

For example, the system engineermay be interested in the technologiesused for the systems to comply withlegal regulations, in the accessibility tothe control settings, and so on. Thesystem engineer is not interested in thesurrounding environment, nor in theweather conditions: s/he is, above all,interested in how the hotel "works in-side" or, more precisely, in its internalquality.

The tour operator takes for grantedthat the hotel works but will care in-stead about the aspects which charac-terize the hotel services at the market:location, price, number of rooms, theaesthetic aspects of the swimming pool,etc. It is interested on how the hotel "isexternally" or, more precisely, in itsexternal quality.

The customer, when s/he uses actu-ally the hotel services, is interested inother aspects. Probably s/he chose thehotel because of its price, location andappearance. However, once in thehotel, these aspects do not interest her/him too much: they are factual evi-dences. Instead, the customer wants thehotel to be adequate for the use s/hewants to make of it. If s/he likes to readin bed, s/he wants there to be sufficientlighting. If s/he gets up late, s/he wantsto be able to have breakfast, and thento find a sun bed at the pool. In otherwords, s/he is no longer interested inthe general characteristics of the hotel:s/he wants it to be adequate for his/herspecific uses. S/he is interested, so tospeak, in the quality in use (or usabil-ity) of the hotel.

Internal quality, external quality andquality in use are general concepts thatcan be applied to any system: hotels,automobiles, or websites. In fact, awebsite can also be considered from aninternal point of view, i.e. from thestructure of the code it is made of, andthis aspect is essential for those whocreated or maintain it. It can also becontemplated from the point of viewof its external specifications, and thisis what the person who evaluate thewebsite is interested in when has todecide if it can go online. Finally, it canbe considered from the user’s point ofview, when s/he uses it in a specificcontext to satisfy personal needs.

These simple examples suggest thatthe definition of a quality model is nota simple matter. In the particular caseof software, the characteristics of aquality model have been discussed formore than thirty years. An ISO stand-ard exists which aims to define pre-cisely the characteristics and sub-char-acteristics of software quality, and theway they can be measured. Even if wehave been inspired by this ISO stand-ard in some aspects, it is not the model

adopted in this article because it is verycomplex, and not particularly adaptedfor websites, which are software sys-tems with very peculiar characteristics.In particular, our model does not looksinto the internal quality (i.e. the inter-nal technical aspects of a website arenot evaluated), but only into the exter-nal quality and the quality in use.

To introduce the quality model wereturn now to the diagram in Figure 1,which shows the actors involved in theproduction and management of awebsite. The first phase is very impor-tant because it is where the objectivesof a website are established; obviouslyits quality can be evaluated only in re-lation to these objectives. Once theobjectives are defined, every phase ofthe work and the actors involved con-tribute to the final quality of the websitefor specific aspects. Figure 3 presentsthe 7 macro-characteristics on whichthe proposed quality model is based(architecture, communication,functionality, content, management,accessibility and usability), shown be-side the activities which contribute themost to each characteristic.

The first characteristic, architec-ture, is about the general structure ofthe website, and the navigation modal-ity which is available to its users. It thiscontext it is applied here exclusivelyto the informative architecture of thewebsite, which is how a website isstructured in relation to its informa-tional contents, and not to its internalsoftware architecture. Internal qualityaspects are not of interest here, sincethey do not deal with the model, asexplained previously. A website hasgood architecture if its page organiza-tions is consistent with its contents andif navigation is easy. This characteris-tic has been indicated in Figure 3 nextto the web design phase, since it is herethat principal decisions are made whichdetermine the general organization ofthe website.

The second characteristic, commu-nication, unites many aspects: the clar-ity of the website for communicatingits message, the consistency with thecorporate image of the website’s owner,the appeal of the graphics, the"relationship" established with the user.The visual design phase above all (but


UPENET

not only) determines thecommunicational quality of thewebsite, as shown in Figure 3.

Functionality is the third charac-teristic considered in the model, andconsists in the set of functions thewebsite makes available for its users.From this point of view, a website is agood one when it provides the user withthe functions s/he requires, all thefunctions work correctly as far as soft-ware is concerned, i.e. is is error-free.To provide a good evaluation of func-tionality, the main functions of thewebsite must be examined in detail andrated for adequacy and correctness.Good functionality derives principallyfrom the correctness of the softwaredevelopment activities, and thereforethis characteristic has been placed inthis area of the figure.

The fourth characteristic is relativeto the quality of the informative con-tents of the website, considering allpoints of view: reliability, updating,comprehension of the contents, and soon. The quality of the contents, as in-dicated in Figure 3, essentially resultsfrom the accuracy of the work done bycontent editors.

The fifth characteristic regards tomanagement and measures the totalquality of the management work of thewebsite during its operations: a qualitywebsite must me managed permanentlyin order to guarantee its correctoperation around the clock. It examinestherefore the work quality of thewebmaster, of the professionals in theweb farm and also of the Internet Ac-cess Provider, since all of them mustdirectly or indirectly guarantee the per-manent operation of the website.

The sixth characteristic is accessi-bility, which covers the aspects whichallow everyone to rapidly access thewebsite without problems. The Web, inthe vision of the World Wide WebConsortium (W3C), intends to guaran-tee universal accessibility (includingusers with disabilities), regardless ofthe type of hardware or software beingused, of the network infrastructure, ofthe language, culture, geographiclocation, ... Therefore accessibility hasmany facets. In order to correctly evalu-ate accessibility, the following aspectsmust be considered: access time to thewebsite (which depends not only onserver efficiency and on theinfrastructures of the available net-

works, but also on the size and "weight"of its pages), easy retrieving ofinformation, and ability to accessinformation through a variety of de-vices, including those which allow forcomprehension by disabled users. Theaccessibility of a website is so a prod-uct of the work done of many actorsand this is the reason why it it has notbeen linked to any particular phase inFigure 3.

Finally, the seventh and last char-acteristic of the model is the usabilityof a website, a term which encompassesall the elements that make a websiteeasy and pleasant to use. From the pointof view of the user, it is perhaps themost important characteristic. It doesnot have its roots in any specific activitybut is the result of the contribution fromall the actors involved in thedevelopment and management of awebsite. Because of this in Figure 3 noparticular activity has been associatedto it: all the activities contribute tocreate usability. It is not independentof the other characteristics: in a certainsense, it summarizes and completesthem all. The other characteristics arenecessary but not sufficient toguarantee a good level of usability. For

Figure 3: Macro-characteristics of The Proposed Quality Model, including Their Main Activities.


UPENET

example, if the access time of a websiteis very long, the usability factor willcertainly be modest. But it does notmean that a website with bad accesstime is usable. Likewise, a set ofcomplete and correct functions guaran-tees good operation of a website, butnot necessarily good usability. Usabil-ity will be tested during access by us-ers in their specific context: each onehas specific traits and needs. Usabilityis also, probably, the most neglectedweb characteristic in practice, becausecreating usable systems is difficult, andrequires paying particular attention tothe users’ needs. The culture of usabil-ity is not adequately diffused yet amongcomputer scientists, and websites todayoften have serious shortcomings fromthis point of view.

Figure 4 represents the proposedquality model. In order to underline theparticular role of usability with respectto other characteristics, it has been rep-resented in the center of the schema.

In the evaluation of the quality of awebsite, it is possible to assign a markto each characteristic, and thereforevisualize in a synthetic form the "qual-ity profile" of a website using a stardiagram. An example is represented inFigure 5, where the merits and defectsof a website can be immediately seen,

even if roughly. It is, so to speak, its"report card". It is particularly advan-tageous for its simplicity, with a scalefrom 0 (which means very poor) to 4(which means very good).

Representation through star dia-grams can be very useful to comparethe quality characteristics of twowebsites: it is enough to overlap theirdiagrams, as in Figure 6.

4 The Sub-Characteristics ofThe Quality ModelAs already observed, the macro-

characteristics of the model have manycomponents. In order to evaluate andassign a mark to each characteristic itis useful to further detail the model andto decompose each characteristic intosimpler sub-characteristics, that can beexamined singularly. In this way, themark attributed to each macro-charac-teristic can be calculated starting fromthe mark attributed to each sub-char-acteristic. In this case there are alsodifferent possible choices. Experienceshows that about twenty sub-character-istics are sufficient to keep the mainquality aspects of a website under con-trol, at least from the point of view ofthe user.

It is a choice which shows in prac-tice to be simple and at the same time

complete enough, as is represented inTable 2. Here, each sub-characteristichas been associated to a simple ques-tion, which is answered during theevaluation of the website.

As seen in the table, the sub-char-acteristics express rather macroscopicproperties of the website, which can befurther decomposed into lower levelproperties. For example, Tables 3 and4 show the decomposition into sub-sub-characteristics of two composite sub-characteristics: Adequacy of theFunctionality, and Graphics of theCommunication.

The complete model cannot be pre-sented here, for space reasons. Overallit has about seventy questions [5],which supply a very detailed guide tothe process of evaluation of a website.These questions are used for every typeof website, independently of their scopeor complexity. If necessary other morespecific questions relative to thespecific applicative domain can beadded to these questions. For example,for e-commerce websites, the question"Are the functions to make the neededtransactions adequate?" can be furtherarticulated as follows:

� Are the functions to select prod-ucts to purchase adequate?

� Are the functions for the inser-tion or modification of payment or ship-ping data adequate?

� Are the functions to confirm theorder adequate?

� Are the functions available forthe user to monitor personal informa-tion and orders adequate?

� Is the security of the transac-tions managed in an adequate manner?

� Are the processes for user reg-istration simple?

� Is the personal information re-quired from users proportional to thetype of transaction?

5 Using The Quality ModelThe quality model which has been

briefly presented in the previous pagescan be used in many ways. It can beuseful for performing a rapid check-upof a website to find the principal areaswhich need improvement, with thepurpose of setting up an action plan. Inthis case, it is not necessary to performFigure 4: Macro-characteristics of Our Quality Model for Websites.


UPENET

Figure 6: Star Diagram Which Allows for The Comparison of Quality Profilesbetween Two or More Websites.

Figure 5: Star Diagram Which Shows The Quality Profile of A Website in AGlimpse.

a detailed analysis of individualsections of the website since it willprobably be heavily restructured: itwould be a useless waste of time andmoney. In other cases, the model canbe used to perform an in-depth evalua-tion of the website, which examines indetail all its principal sections. Forexample, this can be useful followinga restyling of the website, or variousmaintenance interventions in an alreadystable and mature website, to verify thatthe overall quality has not degraded.Sometimes the model can also be usedto compare the website to othercompetitors’ websites, to identifyrespective strong and weak points.

In all cases, it is necessary to ini-tially define our goals with accuracy,to decide how to proceed and how longand how many resources will be usedfor this task. An evaluation done with-out a clear definition of the methodol-ogy would in fact have the risk of pro-ducing results which are not very sig-nificant.

The final result of the evaluation ofa website is usually a document called"evaluation report". Briefly speaking,it contains a description of theobjectives of the work done and the

methodology used. It then presents ananalysis of the quality characteristicsof the website, and the synthesis of theresults obtained: the strong and weakpoints of the website. Finally, the docu-ment concludes with a series of pro-posals for improvement, inspired and

motivated by the analysis. The docu-ment does not need to be very long, inpractice what actually matters is that itcovers all the necessary material to jus-tify and adequately support the im-provement actions proposed.

The definition of the metrics to beused for the evaluation of the differentcharacteristics is an important aspect.Usually in the evaluation of a specificwebsite not all the sub-characteristicsof the model are given the same im-portance. Some characteristics couldeven be considered irrelevant, as forexample the localization of a websitewhich is intended only for a singlecountry. Therefore in function of thewebsite objectives, it is useful to as-sign a weight to each sub-characteris-tic, showing the importance assignedto it in the evaluation. This way, thevalue of each characteristic can be cal-culated as "weighted average" of thevalues assigned to for each sub-char-acteristic.

The example in Table 5 explains themethod. Each sub-characteristic hasbeen assigned a weight, expressed witha number between 0 and 1. The value0 is associated to the sub-characteris-tics irrelevant for the website beinganalyzed, and therefore they do notcontribute to the final value. Instead the


UPENET

Table 2: Characteristics and Sub-characteristics of The Proposed Quality Model.

value 1 is associated to the most im-portant characteristics. For example, inthe evaluation of the Architecture, wehave given the highest importance (1)to Structure and Navigation, and anaverage importance (0.5) to the sitemap. Since the values attributed toStructure, Site map and Navigation arerespectively 3. 2 and 3 (on a scale from0 to 4, like we did for the macro-char-acteristics), the total value (weighted)for the Architecture will be obtained

from the following calculation:

(3 * 1 + 2 * 0.5 + 3 * 1)/2.5 = 2.80

The value is higher than the simplearithmetic mean of the values of thesub-characteristics (2.67). In fact, thesite map, which received a rather lowvalue (2), has a lower weight than theother sub-characteristics.

As seen previously, the overallevaluation of the website can be repre-

sented by a star diagram, like in Figure5, which is relative to the example inTable 5. It can be also worthwhile tosee the evaluation of the sub-charac-teristics with a star diagram. For exam-ple, the diagram in Figure 7 representsthe evaluation in Table 5.

The diagrams in Figures 5 and 7 givea fairly precise vision of the physiognomyof the website, of its strong and weakpoints even if we do not know anythingabout it. They are somewhat like the

1 This sub-characteristic examines multi-language websites, and deals with the correct translation of the informational contents, taking intoaccount not only the linguistic differences between different countries (currency, measurement system, different conventions, etc.).2 This sub-characteristic summarizes essentially three very important aspects which can contribute decisively to the success of a website, andtherefore should be managed carefully: the mnemonics of the URL of a website (Is the name of the website easy to remember?); the rankingin search motors (Is the website easy to find with the most popular search engines?); the popularity of the website (Is the website adequatelyreferenced by other websites?).3 The sub-characteristics for usability have been identified according to the ISO 9241 standard, which defines usability as "the effectiveness,efficiency and satisfaction necessary for users to reach precise objectives in a precise use context". In this definition, effectiveness means "theaccuracy and completeness necessary for users to obtain the desired results"; efficiency it means "the resources spent to obtain this result",and satisfaction is "the comfort and acceptance of the system by the user" [3].

Characteristic Sub-characteristic What we should we ask?

Structure Is the informative structure of the website adequate? Site map Is there a site map of the website which clearly

represents its structure?

Architecture

Navigation Is the navigation of the website adequate? Home page Does the home page clearly communicate the

objectives of the website? Brand image Is the website consistent with the brand image?

Communication

Graphics Are the graphics of the website adequate? Adequacy Are the functions of the website adequate? Functionality

Correctness Are the functions of the website correct? Categorization/labeling Is the information classified in an adequate manner?

Style Is the style of adequate for the website? Information Is the information adequate, pertinent, reliable, and

updated?

Contents

Localization Is the website correctly localized?1 Availability Is the website always on and available? Monitoring Is the use of the website adequately monitored? Updating Is the website constantly updated and improved?

Management

Relationship with users Is the relationship with users adequately supported? Access time Is the access time adequate?

Easiness to be found Is the website easy to find?2 Browser independence Is the website accessible with every browser?

Accessibility

Accessibilty by disabled people

Is the website usable by users with disabilities?

Effectiveness Can the user obtain the desired result in an accurate and complete way?

Efficiency Is the effort requested from the user to obtain the desired result acceptable?

Usability3

User satisfaction Is the website comfortable and well accepted well by the user?


UPENET

Table 3: Decomposition of Sub-characteristic Functional Adequacy into Sub-sub-characteristics.

results of the analysis of a clinical check-up, when the doctor reaches hisconclusions without knowing anythingabout the patient’s life.

In order to demonstrate our claim,let us try to analyze the data in the ex-ample. We can see from Figures 5 and7 that the website being observed isexcellent from a functional point ofview, it has good contents (but it couldbe improved in classification and la-belling), it has good graphics (but thecoherence with the brand could be im-proved). The management is only sat-isfactory: in fact the availability of thewebsite could be improved (perhaps theservers are unstable) and also thequality of the updates (Is there obso-lete information?), and access monitor-

ing tools are not used. The website isnot accessible by the disabled, but thiswas not an objective (the weight attrib-uted to this characteristic is only 0,25).The overall usability could be im-proved (the value of effectiveness, ef-ficiency and user satisfaction is only amediocre 2). This does not seem to becaused by important structural prob-lems (the structure and the navigationare good), but is probably due to a se-ries of microfactors which togethermake the total value lower.

These factors must be identifiedcarefully by analyzing the usabilityproblems which emerged during theevaluation in order to discover whereusers had the most difficulty. It seemsthat these most critical points are likely

concentrated in the labelling (which hasa low evaluation). Perhaps the lowimportance given to the site map (only0,5) should be reconsidered, as it couldbe improved.

These numbers seem to tell us thatwe are dealing with a website with sev-eral positive characteristics, with weakpoints from the point of view of usabil-ity, which nevertheless could beimproved without big structuralchanges or functional additions, andwhose management we definitely needto improve. The brand consistencyproblem must presumably be furtheranalyzed by communication depart-ment of the company. When the finalsynthesis, the evaluators obviouslymust not go forward blindly as in the

Table 4: Decomposition of Sub-characteristic Graphics into Sub-sub-characteristics.

Sub-characteristic

Sub-sub-characteristic

What we should ask

Content management

Are the functions available for website editors for the insertion and updating of the informative contents during its life adequate?

Transactions Are the transaction functions adequate?

Search Are the tools for searching information adequate?

Error management

Are user errors handled adequately?

Adequacy

Are the functions of the site adequate?

Communication Can the user contact the website manager adequately?

Propietario

Placed Image

Propietario

Placed Image


UPENET

example just examined, where we haveonly considered the numbers. Theywill arrive at their conclusions aftercarefully analyzing the website indetail, and having discussed it with asample of users. Even if the valuesattributed by them to the manycharacteristics will be subjective bynature, they will know very well thedefects and merits of the websit. Theresulting diagnosis will therefore berather reliable.

5 ConclusionsThis article has presented, though

very synthetically, a quality model forwebsites, which evaluate both exter-nal quality and quality in use. This

model has been experimented andimproved on over several years,init ially and essentially foreducational purposes, to helpstudents in university courses taughtby the author to develop a globalability of critical analysis. In particu-lar it has been used in order to dorapid check-ups and in-depth evalu-ations of the quality of numerouswebsites of every level of complexity,ranging from simple informationalwebsites to e-commerce websites andlarge portals.

The main merit of the model con-sists in the fact that it is based on avision of the development and man-agement process of a website, as

depicted in Figure 3. In this way theimprovement proposals which arisefrom the evaluation of a website canbe easily sent to the differentcategories of participants involved.Another significant merit is itssimplicity: it can also be used bypeople who do not have particulartechnical competencies about webtechnologies. Furthermore, themethodology can be easily upgraded:it can be used to do quick informalcheck-ups of a website (in a very lim-ited amount of time: a few work days)or in-depth evaluations, which in-volve sophisticated analysis andusability tests.

Numerous examples of the use, by

Table 5: Example of Evaluation of A Website. (The evaluation of each characteristic corresponds to theaverage weight of the evaluation of its sub-characteristics on a scale from 0 to 4.)

Characteristic Mark Weight Weighted value Architecture 2.67 2.50 2.80 Structure 3 1 3 Site map 2 0.5 1 Navigation 3 1 3 Communication 3.00 3.00 3.00 Home page 4 1 4 Brand image 2 1 2 Graphics 3 1 3 Functionality 4.00 2.00 4.00 Adequacy 4 1 4 Correctness 4 1 4 Contents 3.00 4.00 3.00 Categorization/labeling 2 1 2 Style 4 1 4 Information 3 1 3 Localization 3 1 3 Management 2.67 4.00 2.00 Availability 2 1 2 Monitoring 1 1 1 Updating 2 1 2 Relationship with users 3 1 3 Accessibility 2.50 3.00 2.92 Access time 3 1 3 Easiness to be found 4 1 4 Browser independence 2 0.75 0.25 Accessibility by disabled people 1 0.25 0.25 Usability 2.00 3.00 2.00 Effectiveness 2 1 2 Efficiency 2 1 2 Satisfaction 2 1 2


UPENET

computer science university students,of this methodology for the evaluationof Italian websites are available at<http://www.ilcheckupdeisitiweb.it>, awebsite managed by the author.

Translation by Julia Weekes

References[1] S. Krug. Don’t Make Me Think – Acommon sense approach to WebUsability, Que, 2000.[2] ISO/IEC 9126: Information Tech-

nology – Software Product Evalu-ation – Quality Characteristics andGuidelines for their Use. 1991 and2000.

[3] ISO 9241, Ergonomic requirementsfor office work with visual displayterminals.

[4] J. Nielsen. Designing Web Usabil-ity, New Riders Publishing, 2000

[5] R. Polillo. Il check-up dei siti web,Apogeo, Milano, 2004.

[6] L. Rosenfeld, P. Morville. Informa-tion Architecture for the WorldWide Web, O’Reilly & Associates,1998 and 2002.

Figure 7: Star Diagram of The Values Attributed to The Sub-Characteristics ofThe Example in Table 5.

Monograph: Computing Omnipresence (published … The Open Source Software vs. Proprietary Software...

Documents

Transcript of Monograph: Computing Omnipresence (published … The Open Source Software vs. Proprietary Software...