Rochester Institute of Technology
RIT Scholar Works
Theses
Thesis/Dissertation Collections
2004
Rationale for user oriented design technique
selection
Susan M. Saeger
Follow this and additional works at:
http://scholarworks.rit.edu/theses
This Thesis is brought to you for free and open access by the Thesis/Dissertation Collections at RIT Scholar Works. It has been accepted for inclusion in Theses by an authorized administrator of RIT Scholar Works. For more information, please [email protected].
Recommended Citation
Rationale
for
UserOrientedDesign Technique Selection
By
Susan M. Saeger
Thesissubmittedinpartialfulfillmentofthe requirementsforthe degreeofMasterofScience in Information
Technology
Rochester Instituteof
Technology
B. Thomas Golisano College
of
Student Name:
Project Title:
Name
Rochester Institute of Technology
B. Thomas Golisano College
of
Computing and Information Sciences
Master of Science in Information Technology
Thesis Approval Form
Susan M. Saeger
Rationale
for
User Oriented Design Technique Selection
Thesis Committee
Signature
Date
=..!Ev~el..J-!.'yn~R~oz=an~ski~· p~hD=---_E_v_e_l~y_n_R_o_z_a_n_s_k_i _---=-.j_-:~d
g
---0
y
Chair
Anne Haake PhD
Committee Member
Deborah Coleman
Committee Member
Anne Haake
Thesis Reproduction Permission Form
Rochester Institute of Technology
B. Thomas Golisano College
of
Computing and Information Sciences
Master of Science in Information Technology
Rationale
for
User Oriented Design Technique Selection
I, Susan Saeger, hereby grant permission to the Wallace Library of the
Rochester Institute of Technology to reproduce my thesis in whole or in part.
Any reproduction must not be for commercial use or profit.
Acknowledgments
I wouldliketoextendedathankyoutomy wonderfulparents, TheresaandJoe
Wanchissen,
whohave alwaysencouraged meto "pushthepencilharder"
Also,
thankyouto
brothers,
JoeandSteve forlistening
ateveryfamily
eventaboutmyeducational status.Iwouldalsoliketo thankEdandJackie Saegernotonly for
believing
in me,but for allowing metovisitthemwhenIneededtorecharge and reestablishmyperspective.I am gratefulforall ofmy friendsandmy husband's friends whohave become
myfriends. You have allbeenso understandingofmyschedule when school hadto
come first. Atthesametime youallknewwhenIneededto takeabreakandlivealittle. Youarethepeople whohelpedto
keep
mebalanced. You are all alittleresponsiblefortheinitial ideaofthispaperbecause itcrystallizedasJeffandIwere
training
forthehalf marathon you all organizedin 2003.My
co-workers and managers alsodeserveathankyouforlistening
andencouragingme notonly
during
thecreation ofthispaper,but forthepast6 years. Allof you who have beenthere fromthestartknowthechallengesIfacedand your supporthas been greatlyappreciated.Thankyoutoall ofmyprofessorswhohavecontributedtomyeducational experience. Mostof allIwouldliketo thankmycommittee members andespecially Evelyn Rozanski PhD. Shewastheperson who established andfostered myexcitement aboutthe topicofthis paper. She isnotonly a wonderfulprofessor,but also a mentor and afriend. Iamgoingto miss our
bi-monthly
meetingsthatincludedtopicsfar beyond whatis presentedinthis paper.Mostofall, averyspecial thankyou goes tomy husbandwhohas beentherefor methroughitallfromthevery beginning. Thank youforalways
believing
inme andhelping
tokeep
my life appropriately balanced. Thank youforknowing
whenIneededtoTableofContents
1. Introduction 9
2. Statementoftheproblem 12
3. Hypothesis 13
4. Definitions 13
5. Background 14
6.
Participatory
Design(PD)
156.1
History
156.2 Definition 16
6.3 Techniqueand processsummary 17
6.4 Projectteam type 20
6.5 Application domain 20
6.6 Integration
feasibility
216.7 Advantages 21
6.8 Expectations 22
7. Scenario-Based Design
(SBD)
237.1
History
237.2 Definition 24
7.3 Techniqueandprocess summary 24
7.4 Projectteamtype 26
7.5 Application domain 26
7.6 Integration
feasibility
267.7 Advantages 26
7.8 Expectations 26
8. Contextual Design
(CD)
268.1
History
268.2 Definition 26
8.3 Techniqueandprocesssummary 26
8.4 Projectteam type 26
8.5 Application domain 26
8.6 Integration
feasibility
268.8 Expectations 26
9. UserCentered Design
(UCD)
269.1
History
269.2 Definition 26
9.3 Techniques and processsummary 26
9.4 Projectteam type 26
9.5 Application domain 26
9.6 Integration
feasibility
269.7 Advantages 26
9.8 Expectations 26
10. Performance Centered Design
(PCD)
2610.1
History
2610.2 Definition 26
10.3 Technique and process summary 26
10.4 Projectteam type 26
10.5 Applicationdomain 26
10.6 Integration
feasibility
2610.7 Advantages 26
10.8 Expectations 26
11.
Learning
Centered Design(LCD)
2611.1
History
2611.2 Definition 26
11.3 Techniqueandprocess summary 26
11.4 Projectteamtype 26
11.5 Application domain 26
11.6 Integration
feasibility
2611.7 Advantages 26
11.8 Expectations 26
12.
Usability
Engineering
(UE)
2612.1
History
2612.2 Definition 26
12.3 Technique andprocess summary 26
12.5 Application domain 26
12.6 Integration
feasibility
2612.7 Advantages 26
12.8 Expectations 26
13. Fictional Case
Study
Purpose 2613.1 Formulation ofProblem Statement 26
13.2 Background Scenario 26
13.3 Criteria fortechnique selection 26
13.4 AnalysisoftheCCTA Project 26
13.5 CCTA Project Goal andScope 26
13.6 CCTA Project Timeline 26
13.7 CCTA Project Team SizeandExperience 26
13.8 CCTA Project Team Type 26
13.9 CCTA Project Team Location 26
13.10 CCTA Project Team Dynamics and organizational culture 26
13.11 User Oriented Design approach recommendation 26
13.12
Participatory
Design(PD)
elimination explanation 26 13.13 Scenario Based Design(SBD)
elimination explanation 2613.14 Performance Centered Design
(PCD)
elimination explanation 26 13.15Learning
Centered Design(LCD)
eliminationexplanation 26 13.16 User Oriented Design(UOD)
techniquerecommendation analysis 2613.17 Analysistechniquesincluded 26
13.18 Analysistechniquesexcluded 26
13.19 Designtechniquesincluded 26
13.20 Designtechniquesexcluded 26
13.21
Testing
techniquesincluded 2613.22
Testing
techniquesexcluded 2613.23 Additional deliverables 26
13.24 Expectedproject outcome 26
13.25 Otherconsiderationsthat were outsidethescope ofthis case study 26
14. Conclusion 26
15. AppendixA
-Methodcomparison
by
technique 261. Introduction
Intoday'scompetitiveenvironment,companies cannot affordtoproduce unusable products. This is particularlytrueforcompaniesthatproduce software, astheirsole revenue generator. Consumers choose software productsthatmaketasks easier. This meansthatusability and usefulness of a software product are nolongerrequirementsthat
are "nicetohave", butmustbecore requirements of software applications.
Achieving
theserequirements is a complextask thatrequiresintense analysis and commitmentfrom several areas of an organization. Partofthereasonthis taskis sodifficult is becauseof thewide range of perspectivesthatinfluencethe creation ofsoftware products.Typically,
the creation of a productinvolvesstakeholdersfrom all areas of anorganizationincluding:
Marketing,
SoftwareEngineers,
HardwareEngineers,
Quality
AssuranceSpecialists,
ImplementationSpecialists,
ProjectManagement,
andDirectors. Whenpeople with differentperspectives cometogethertomove aproductfroma concept toreality, it is inevitable thatall parties must compromisetoachieve success. Thefirst step toachieving success istounderstandtheoverall organizational goals; thesecond step is understanding theconsumer or end user oftheproduct.As achieving aclear organizational goal isnotthefocus
here,
theassumptionis that theorganization already hasaclearly identified goal.Rather,
thefocusis to understandhowtoselectthe appropriateuserorienteddesignmethodand techniques toachievetheorganization's strategic objective andthespecific project's goals.
Userorienteddesign methods aremultidisciplinary because
they
considerthepsychology, ergonomics, sociology, engineering, and visualdesign factors affecting productcreation. Auserorienteddesignperspectivefocuseson theend user andthe
interaction betweenend user and product. Thegoal of userorienteddesignmethods isto
add value
by incorporating
thehuman aspectinto productdesign. Becauseofthis humannatureperspective, userorienteddesign methodscanbeappliedtomany differenttypes
of products and systems.
varietyof reasons. Thequestions
they
askedincluded:"Which user-centered design
(UCD)
methods are most widely used andwhy?What are the benefits and weaknesses of each method in the
eyes ofpractitioners?
What are the organizational impacts of UCD and what measures are in place to assess
progress?"
(Vredenburg, Mao,
Smith,
Carvey
2002p.471)
(Note: theuse ofthe termuser-centereddesign
by
theauthorsisequivalentto the termuser orienteddesignmethodsthatare referredto inthispaper.)This studyrevealedthatUCD isnot well-establishedintheindustry.
Recently
Rosenbaumcompleted another study."Rosenbaum et al.
(2000)
surveyed 134 CHI professionals with afocus on the contribution of organizational approaches and UCD methodsto strategic usability. It wasfoundthatmajor obstaclesto creating greater strategic impact include resource constraints, which was mentioned
by
28.6% oftherespondents. Resistance to user-centered design or usability, lack of knowledge aboutusability were also stated as obstacles.
However,
partnering withmarketing was identified as a very effective
approach"
(Vredenburg, Mao, Smith,
Carvey
2002p.472)
to UCD.Hudsonconducted an e-mail survey in 2000 and 102 usabilitypractitioners
revealedthat the mostcommonlyusedtechniques are: informal usability testing,user analysis,userprofiling, evaluating existingsystems,
low-fidelity
prototyping, heuristicevaluation, task
identification,
navigationdesign,
and scenario-baseddesign(Vredenburg, Mao, Smith,
Carvey
2002). Asimilar 10-questionweb survey,conductedin 2001
by
Gunther, Janis,
andButlerthatinvolved 100 usabilitypractitionersfoundthat 39%oftherespondentsbelieved usabilitytesting
tobethemostaccuratein gauging success(Vredenburg,
Mao, Smith,
Carvey
2002). Thestatisticsindicate thattheselection andimplementation ofthe appropriateuser orienteddesignmethodis not simple andthatthereis agap in knowledgeabout user orienteddesignmethods.
results
(Vredenburg, Mao, Smith,
Carvey
2002). Theresultsfoundthat32%ofrespondents were notsureif UCDmethodshad helpedsave product developmentcosts;
only 24%believedthatUCDmethodsactually saved productdevelopmentcosts and; 44% thoughtthatUCD increasedtheproductdevelopmentcost. Similarresults were
foundwhenrespondentswere asked abouttheproductdevelopment time.
Vredenburg,
Mao, Smith,
andCarey
arguethat theresults areinconclusive becausetherespondents maynothavelooked atthe"big
pictureincluding
service cost andredesign"
(Vredenburg, Mao,
Smith,
Carvey
2002p. 474).Nevertheless,
theVredenburg, Mao,
Smith,
andCarey
studywasconsistent withHundson's informale-mailsurveytakenin 2000. Both studiesindicatethatusability testing,prototyping, andheuristicevaluations arethemostwidelyusedlowcost methods.Vredenburg, Mao,
SmithCarey
concludedthat:"UCD method adoption is quite uneven across different organizations.
UCD staff in many organizations, 41% of
[the]
sample, is centralized, andonly 15% oftheorganizationshave completely decentralized UCDstaff.The average spending on UCD constitutes about 19% of the total projectbudget.
There is a lack of measurement ofUCD effectiveness and any common evaluation criteria acrosstheindustry.
A multidisciplinary approach to UCD appears to be closely related to perceived UCD effectiveness, although practitioners were not always clear about what constitutedmultidisciplinary. UCD was perceived to have higher impact when there were two or more UCD specialists on the project team compared withonlyone.
Cost-benefittradeoffsplay a major rolein theadoptionofUCD
methods"
(Vredenburg., Mao,
Smith,
Carvey
2002p. 477-78).Productively implementing
userorienteddesignmethods requires awarenessof theorganizational structure.Vredenburg, Mao, Smith,
andCarey
suggestedthatthe organizations usedinthe studies maynothave properly alignedits staff(Vredenburg,
Mao, Smith,
Carvey
2002). "Respondentswere askedtocharacterize theorganization of UCDstaff. In 41%ofthe companies,UCDstaff were centralizedinan organizational unit, 15%decentralized,
34% mixed,and 10% unclear"Carvey
p.474).Understanding
themethodsandtechniquesprovidesinsightabouthow to appropriatelyalign staffmembers.Theinformationprovidedinthispaperwillbegin toclosethegap inthe
understandingofthevarietyof methods andtheirtechniques thatprovide anorganization
thetools to movetowardsa user-centric designapproach. The intent istoprovide a
comprehensive comparison of user orienteddesignmethods,and guidelines formethod
andtechnique selections. Itprovidesthereasoning behindthe selection oftechniques ratherthan randommethod selection. The focusofthis paper willbeon selectinga user
orienteddesignmethodforsoftware applicationdevelopment.
By
providingacentrallocation fortheinformationrequired whenselecting user orienteddesignmethods and
techniques,thispaper will closethegapthatcurrentlyexistsItheuser orienteddesign industry.
2. Statementoftheproblem
Userorienteddesign's broadnaturehasprovidedtheopportunity forpractitioners tocombinetechniquesin a uniquefashion
thereby
creating several similar, yetdifferent approachestouser-centric software productdevelopment. As practitionershavesuccessfullycombinedtechniques thathavecreatedperceived successful software application
design,
thepractitionershaveclassifiedtheselectedtechniques with methodologynames. Thepractitionersthathavemovedinto consultingroles align themselveswith themethodthatcreatedtheirinitialsuccess andthen specializeinthat method. Thepervasive userorienteddesignmethodsthatconsultantshavealignedthemselves withincludethefollowing:
Participatory
Design(PD)
Scenario Based Design(SBD)
Contextual Design(CD)
User Centered Design(UCD)
Performance Centered Design
(PCD)
Learning
Centered Design(LCD)
Usability
Engineering
(UE)
Sinceconsultants alignthemselveswith a method it isproblematic foran
sustainingprocess towardsuser-orienteddesign inthe organization.
Speculationcanbemadethat the selection ofauserorienteddesignapproach
within anorganization happensinconjunction with an immediateneed of ahighprofile project. Furtherspeculation canbemadethatfororganizations newto a user oriented
designapproach looktoconsultantsforsupport. Anotherassumptionisthattheinitial
selectionof a user orienteddesign methodonly includesa review oftheprominent consultantandthemethods
they
offer. Thistypeof selectionmayresolvetheimmediateunderstandingoftheprojectobjectives,but maynot resultin asustaininguseroriented
design approach.
3. Hypothesis
If knowledge about proper user orienteddesigntechniqueselectionis availablethen an organizationwillbeempoweredtocustomizea user orienteddesign approachfor successful integration into itsunique softwaredevelopmentproject requirements
resulting in a sustainable user orienteddesign approach. Asustainable user oriented
design approach willhavea positive strategic effectfortheorganization.
Theresultofthe analysis containedinthispaperis a set ofguidelines for
selectinga user orienteddesign technique. A fictionalcasestudy is usedtoillustrate the application ofthe guidelines. Included inthecase study isan analysis ofthemethods and theirtechniques; whereby, thetechniqueshave been
logically
groupedby
the threemajor stages of a software application projectthosebeing: requirements,design,
andtesting. Inaddition, thedeliverablesprovide
by
each method arediscussed. 4. DefinitionsProjecttypes, cross-functionalteams, andtraditional software developmentareterms
usedthroughout thisdiscussionthatrequires adefinitionto groundthebasesofthe analysis.
Thegeneralreference toprojecttypesis basedon theproject objective andincludes
the
following
list (Olson 2004): NewproductReengineering
projectMaintenanceto anexistingproduct
.
Redesigning
an existingproductTransitiontoa new platform
Also,
throughoutthediscussion referenceis madetocross-functionalteams. Cross-functional teams are composed of peoplefromvariousfunctional areas whose competencies are essentialfor achievinga successful outcome. Theseteams includeemployees fromalllevelsoftheorganizationandmay includemembersfromoutsidethe
organizationincluding:
key
customers and consultants.Lastly,
SeffahandMetzker(2004)
definethe term traditional softwaredevelopment as
technology
anddeveloper driven with theprimary focus onfunctionalrequirements with littleregardforend users workflow or end users environments.
5. Background
Atfirstglance,thedifferentuser orienteddesignmethodologies seemto sharea common
theme:
keeping
theuserin mind whendesigning
andcreatingsoftware applications.However,
not all methods are equal. Thesolutionto theproblemmaynotbeas simple asthe selection of amethod,butrathertheselection ofthepropertechniques fromthe differentmethods.
Selecting
therightmethod requiresanalysis ofthetechniques withinthemethod. Itrequires consideration of eachtechnique's abilityto facilitatethecurrent projectgoal alongsidethe strategicdirectionoftheorganization.
This discussionof user orienteddesignmethodologies followsthetimeline ofthe
methods emerging intotheindustry. The discussion beginswiththe
Participatory
Designmethod. Someofthemethod's namesmay have been coined earlierthen the
developmentoftheactual method andmay haverootsthatgofurther back in
history
however,
thiswas notfactored intothe timeline. Thechartin Figure 1 representsthe timeline ofthe user orienteddesign methods. Noticeseveraloftheuser orienteddesignmethods emergedin 1989. Theorder ofthemethodsemerging in 1989 is basedonthe
analysis ofthereferences totechniqueswithin apreviouslydefinedmethod. Theanalysis revealsthat eachuser oriented designmethodborrowsconcepts andtechniques fromthe
previous methods and sometimesrenamesthetechniques.
Therefore,
some ofthe overlappingtechniques aremerelyuniqueby
thereferenced name. ThechartinUseroriented
designmethod
(UOD)
Approximated initial availability
Personwho originated
Participatory
design(PD)
1960 Christina FloydKristen Nygaard Olav Terje Bergo
Scenariobased
design(SBD)
1980 Ivar JacabsJohn Carroll
Mary
Beth RossonContextualdesign
(CD)
1988 Hugh BeyerKaren Holtzblatt
Usercentereddesign
(UCD)
1989 Don NormanDr. Steven Draper Performancecentereddesign
(PCD)
1989 GloriaGery
Barry
RaybouldLearning
centereddesign(LCD)
1989 ElliotSoloway
Usability Engineering
(UE)
1993 Jacob NielsonDeborah Mayhew Figure 1- UOD Time Line
6.
Participatory
Design(PD)
6.1History
Participatory
Design(PD)
began in Scandinaviaaround 1960 fromcollaborationbetweentheNorwegianEmployers'
Federation
(NAF)
andtheNorwegian FederationofTrade Unions. Oneoftheoutcomes ofthiscollaboration effort was "arevisedWorker ProtectionandWorking
Environment Act[AML, 77;
S0rensen,
92]"(Levinger 1998). Thisact requiredthatend users participatein thedesignandbe sufficientlytrained touse new systems. Threemajor projects are creditedfor creatingthefoundation ofPD.They
are:The Norwegian Iron andMetalWorkers'
Union
(NJMF)
conducted from 1970tol973. Theirobjective wastoincludeworkersinnewtechnology
plans.. The Swedish DEMOS project(DEMOkratiske
Styringssystemer)
was conductedfrom 1975to 1979. This project confirmedthat theinequalities betweenworkersandmanagementwere particularly unbalanced when
introducing
new technology. Theoutcome ofthisproject was aninstitutionalizedconflict managementprocess (Levinger 1998).
. TheDanish DUEproject
influence ontheuse of computer systems. Theprojectcontributedto theinitiation of aprofessionalcurriculum and research programin systems development (Levinger 1998).
Thecommon themeofthesethreeprojects was todemocratizetheunityoflaborand
management when
introducing
newtechnologies. It iscleartoseethat"PD began inan explicitlypoliticalcontext,as part oftheScandinavianworkplacedemocracy
movement"
(Muller 2002p. 1). Theoutcome ofthese threeprojects was thebases for PD whichis also referredtoasthe Collective Resource Approach.
Themore recentdevelopments in PDresultedfromthe first PDconferencethat tookplaceinl990 in
Seattle,
Washington. Theparticipantsincluded "prominent leaders from Scandinaviawhohadpioneeredparticipatory approachestocomputer systems development inthe70's and80's."
Afterthisconferencefuture conferenceswereheld
everytwo yearstofurther
develop
anddefine PD (Levinger 1998).There is nostrongevidencethatone personisresponsibleforthe developmentof thePD approach. Christine Floyd isthemainEuropeanfigure given creditfor
bringing
PDout ofScandinavia (Bouvin 2004).However,
therootsoftheCollective Resource Approach are generally associated with theeffortsinNorway by
Kristen Nygaardand Olav Terje Bergo in theearly 1970s (Levinger 1998).6.2 Definition
Defining
PD isnoteasybecause,
similartoother userorienteddesignmethods, practitioners vary intheirperspectives,backgrounds,
and areas of concern. Themore recentPD definitionsare asfollows:.
"Participatory
Design(PD)
isa set ofdiverseways ofthinking, planning,and actingthroughwhich people maketheir work, technologies,and socialinstitutionsmore responsive tohuman needs"(TriggandClement2000).
.
Participatory
Design(PD)
is "anapproachtothe assessment,design,
anddevelopmentoftechnological andorganizational systemsthat places a premium on theactiveinvolvementof workplace practitioners(usually
potential orcurrentusers ofthesystem)in design anddecision-making
processes"(Trigg
andClement 2000). ."Participatory
design(PD)
isasetoftheories, practices, and studies relatedtoend-users asfull participantsinactivitiesleading
tosoftware andhardwarecomputer products andcomputer-basedPD isnotnecessarilya"unified
ideology
ormethodology but is a techniquethatcanbeembraced. It is an approachtowardscomputer systemdesign inwhich thepeopledestinedtouse thesystemplaya criticalroleindesigning
it"
(SchulerandNamioka 1993p. xi).
Althoughthere isno singledefinitionfor
PD,
thecommon theme among practitioners whousePD isthat theend user must participatethroughout theentire productdevelopmentprocess. This includesparticipationin preliminary discussionsbetween stakeholders and managers
(Trigg
andClement 2000). PDpractitionersunderstandthat the organizational cultureandthe setting inwhich workis done impacts
design,
andtherefore,PDemphasizestheimportanceofspendingtimewith end users who aredoing
theworkintheirown environment. PDpractitioners alsohave learnedthattheend users aretheprime source ofinnovationand can greatlycontributeto generatingdesign ideas. In addition,PDpractitionerslooktoresolve problemsthat extendbeyond
technology
and ofteninclude theprocess and arrangement of people.6.3 Techniqueand processsummary
According
toB0dker,
Br0nbaekandKyng,
PD is aperspectiveand atechniquenotnecessarily adesignmethod(SchulerandNamioka 1993p. 158). Schuler further
explainsPD as a perspective and notnecessarilyablueprintforadesignprocess. The
perspective ofPD istoachievethedevelopmentof a usableand stress-free software
design. The use ofappropriateinstructions alongwiththeavailabilityoftheappropriate
functions tosupportthe workflow aretwoelementsthatcontributetothe stress-free nature ofa software design (SchulerandNamioka 1993 p. 6-7). PDprovidestheproper techniques toachievethis goal.
Although PD doesnothavea specific process,thetechniques lendthemselvesto theanalysis anddesignphase oftheprojectlifecycle.
During
theanalysis phasedesignteammembers, learnabouttheworkthroughworkplace visitswhere theend users participatein interviews or workdemonstrations. Teammembersmeet afterthe
interviews anddemonstrations tocompile what
they
have learned. PDusesthe scenario, contextualinquiry,
storiesand prototypetechniques all of whichhave been borrowedby
other user orienteddesign methods.
Thetechniquesthatmake PDuniqueareworkshops and organizationalgames.
phase. PD specificallyuses aworkshopcalledthefutureworkshop. Thefuture workshoptechniquewas created
by
RobertJung
andNorbert Miller(1987)
toincludecitizens inthedecisionmakingprocess intownplanning and other political venues. This techniquerequires thattheparticipants write short statements oftheirideas on paper and thenplacethem on awallfortheentireteam tosee. Theoutcomeofthefuture workshop
resembles an affinity diagram. The future workshopthen goesthroughthreephases, whichare; critiquephase,
fantasy
phase,andimplementationphase. Thepurpose ofthecritique phaseistodefine specificissuesabout current work practices. Theteamthen
moves intothe
fantasy
phase. Thepurposeofthefantasy
phase istodefineanideal futuremodel. Theparticipants generate 'what if statementsthatwould solve currentissues. The lastphaseis theimplementationphase. The implementation phase alignsthe ideasgeneratedinthe
fantasy
phase with reality.During
this phase, the teambeginsto designa realistic solution. The use ofthe termimplementationhere isnotthemovementofthesoftware application to theendusers,butrathertheimplementation ofideas intoa design.
Thetechniqueof
'design-by-playing
games'
alsobridges theanalysis workinto designwork.
Ehn, Sjogren, B0dker,
Gr0mbaekandKyng
developedtheideaof'design-by-playing
games'
because ofthe abilityofthesetechniques toinfluence participatory designmethods (SchulerandNamioka 1993 p. 168). Thegames supportthecreation of
alternative work organizations
by
enactingthecurrent routine andthendiscussing
problems. It includes thediscussionabouttheproblemsthata newalternative maycreate.
Using
games creates alessemotional environmentfor discussion. Gameskeep
theworkentertainingand atthesametime,the teamfocuseson theworkflow. These techniquesallow formembers unfamiliar withthedesign discipline aneasy wayto contribute. Mullerexplainsthat thegamesremove "theconventional authorityofthe software professionals [andreplacesit]
with a sharedinterpretation basedoncontributionsfrom multipledisciplinesand
perspectives"
(Muller 2002p. 18). Thistype ofdiscussion allowstheanalysis anddesignprocesstomove ataquicker pace and produces betterend products becausepeoplelistentoeach other's needs andthe
techniques thatcreateasafeand productive atmosphereforpeoplefrom different perspectivestodiscuss conflictingopinions. ThechartinFigure 4 summarizesthese games.
Game Name Description
Organizational Game Acharades-like card gamethatfacilitatesthediscussionof current workflow problems andto achieve a suitable solution. Specification Game A scenario-based gamebasedon a set of 'situationcards',each
of which describeda workplace situation. Thedesignteam takes turns
drawing
acard andleading
thediscussion ofthe work situation describedonthecard.Layout Kit A game offloor-plans and equipmentlayout. This illustrates the workers'
perspective ofhowtheshop floorshouldbe redesigned.
Desktop Publishing
Game A story-board gamethatillustrates and annotates on poster size paperthecomponentsof work.Figure 4 Design Game
Summary
(Muller 2002p.17)
Stories are anothertechniquePDembraces.
Stories,
just likeworkshops andgametechniques,
bring
various perspectivestogetherduring
theanalysisphase. "Stories in participatoryworkmay function in atleastthreeways: triggersforconversation,analysis,and
feedback"
(Muller 2002p. 11). Ifan end usertells thestory,itprovidesthe
teamwithbackground informationtounderstand whatfunctionsandfeaturestheproduct must provide.
Alternatively,
ifthe story istoldby
amember ofthedesign team, it isusedtopresenttheirconcept of what adesignedproduct will
do,
how itwill beused, and what changes will occur as a result.Drama isanothertechnique thatsupportstheanalysisphaseinthePDmethod. "Dramaprovides anotherwayto tell stories
-in theformoftheatre or of video"(Muller 2002p. 14). The drama mayrepresent a currentsituation or afuture situation.
However,
it issometimes difficulttodefinewhether agivendramadepictsthepresent orthefuture because the twoare oftenmixed.Once the teamhas donethe analysis,
they
canmoveintothedesignphase withtheuse of prototypes. Prototypes arethe mostcommonlysharedtechniqueacross all user orienteddesign methodologies.
However,
B0dkerandGr0nbaek(1991)
introducedtheprototypinginthat traditional prototypingapproachesmainlytake theperspectiveofthe
analystor
designers,
whoconductinvestigations intheend user'senvironment,develop
prototypeson their own, andthen sharetheresults with thesoftware engineers and
eventuallywith theend users
(B0dker,
Gr0nbaekandKyng
p. 170). Cooperativeprototypingmeans thattheend usersactuallycreatetheprototype alongsidetheother
teammembers.
6.4 Projectteam type
PDrequires morethana cross-functional team. PDhas been successfulwhere unions are
popularand when
'Workers'
Billof
Rights'
have beenestablished. The hierarchical
structurethat themajorityofU.S.companies use isnotconducivefor PD. PDwould
require majorrestructuring formostU.S. companies. U.S.software development
companies would needto integratesoftware engineers withtherest oftheorganization,
and allow software engineersto interact
directly
withtheend users.6.5 Application domain
The PDmethod wasprimarily designedtofinetuneexistingsystems. PD is successful
on projects with well-defined scopes and unlimitedaccessibilityto theend users. PD asa
methoddoesnot work wellforgeneral use software creationorinsituations wherethe
end usergroup is largeanddiverse. GrudinandPruittexplainthat"cooperative design
techniques canbeeffectivein in-houseor customerdevelopmentcontextsbutareless
effective incommercial productor package software
development"
(Grudin andPruitt
2002 p. 1). Theproblem withPD formass-market softwaredevelopmentproducts isthat
it isnot reasonable tohaveend user representativesforeachtype of usergroup
participatein the analysis, design andevaluations. Thetimelines do not allowforthis
typeof participation andthecost would not provide enoughbenefit. This isnottosay
thatPDas atechniqueis worthless,just as an end-to-end methodit isnot reasonablefor
mosttypes of software productdevelopment.
Inaddition, the 'Workers' Bill of
Rights',
thatare establishedby
unionsmaycontribute to thesuccess ofPD. This 'Workers' Bill of
Rights'
setstheguidelines and
keepsthedesigners andsoftware engineers accountableto theend users. It is
questionableif PD wouldbesuccessfulin an environment were such agreements were
6.6 Integration
feasibility
PD isanapproachthatrequiresa cultural change within an organization whichtodate haspreventedthis methodfrom seamlessly
integrating
intothesoftwaredevelopment industry. Thistype ofchangerequires changingan organization's culture.Grudin,
aUsability
Engineer atMicrosoft,
explainsthat,"participatory
designismoredifficultto accomplish outsideoftheScandinaviancountriesbecauseofdifferencesin legislative environment, workplaceunionization, and scale andfragmentation of softwaredevelopmentorganizational
models"
(Muller 2002p. 229).
Implementing
atruePDapproach would require an organizationtoallow end usersto
fully
participateon projectteams andbetreatedas equalteammembers and notjustas peripheralteammembers who areinvolved in requirementsgatheringandthenare not spoken tounless
compromisesto therequirements are required.
Changing
thewayan organizationthinks about end user participationdoesnothappenovernight. Thetraditional one-directional approachtosoftware productdevelopmentwould needtobecomeatwo-way
discussion werethe analysts,designers,
software engineers and end users all participatein thediscussionofthesolution (Muller 2002p. 6). ThistypeofdiscussionrequiresITto removethe walls and
truly
listento theend usersthroughout theprojectlifecycle. These meansthatunlesssoftwaredevelopmentorganizationsarewillingtomodifythelevelof priority forend userinvolvementand arewillingtotruly
listentotheend users needsthebestthePDmethod canhope for istobeused as atechniqueandtocontinuetobe integratedwith other user orienteddesignmethods.
6.7 Advantages
PDgoesbeyondother methodstoinclude end usersas afull-fledgedteammember. It strivesfor designersand end userstohave a practicalunderstandingof each other. Practical understanding differs fromprepositionalknowledge. Practical understanding
"is understandingthatcomesfromthe practical experience of
doing
somethingandthe recall tomind of earlierexperiences"
(Ehn 1993). Prepositional knowledge is
being
able tousedefinitionstoexplainsomething thathasneveractuallybeenexperienced(StanfordUniversity). PDencouragesdesign
by doing
thusminimizingthelanguageAnotheradvantageofPD isthatitencouragesdecentralizeddecision making
because end users participateintheentire designprocess. Endusers are notonly
observedand
interviewed,
but actuallyparticipateinthecreation ofthedesign.Thegoal ofPD istounderstand organizational changethat will occur andmay
includerestructuringtheend users workflow when
introducing
new ordifferenttechnology
in organizational structures (GrudinandPruitt2002p. 99). PDminimizesthenumberof surprisesthatthenew
technology
maycreateinthe workflow. This is becausePDencouragesa mix of nonlinguistic andlinguistic design artifacts.
Using
nonlinguistic techniquessuch as mockups and prototypes helpsteammembers see a pictureandminimizestheamount ofinterpretation.
Using
linguistictoolssuchas scenarios also allowsforallteammembers toparticipatein and understandtheend users process priortodesigning. PDconsidersmorethan theend usersinasilo,butrathertheendusers
within theorganizational structure.
In addition,PDencourages and supportsearlyproduct adoptionfromthe end user
community. There isevidencethatanincrease inend userinvolvementthrough
participationin design activities canleadtoincreasedwillingnesstouse theproduct
whenit is implemented (Muller 2002).
Thus,
PD canbevaluable notonly in shapingtechnology
to theend user'sneeds, butalsoin shapingtheenduser's willingnesstotry
anew
technology
(Muller 2002p. 299).Theultimate advantage ofPD isthat thenew software product will not only be
useable and usefulbuttheend users willhavea greaterunderstandingofhow the
applicationactuallyworks.
6.8 Expectations
Presentandfuturesoftwareisexpectedtobemoreinteractiveand unobtrusiveto theend
userthanearlier softwareapplications.
Therefore,
new software designcannot excludefrequentparticipation fromtheend users. PDclaims other user orienteddesignmethods
donot
truly
include end users. The reality isthat theother methods includeend users aspost adhocmembersof ateam thatsimplyreview and sign off on requirements.
However,
embracing PDwillbe difficult intheUnitedStates primarily becauseofcomputer products shouldtakepartin thedecisionsthataffectthesystem andtheway it
is designedand
used"
(Greenbaum 1993p. 28). Although theU.S.cultureisconsidered
democratic,
whenitcomes tosoftwaredesignanddevelopment,
afewpeopledictatetheoutcome. In
theory
software engineers agreethatend users shouldbeinvolved, however,
inpractice software engineersoftenfeelthatend users makeimpossiblerequests thatcannotbeaccommodatedin the timeframeandbudgetestablished
by
stakeholders."Many
countrieshavelegislatedthatrepresentatives ofdifferent interests be involved inprojects; elsewhere,management andlabor haveagreedtoconsultationoverany proposed changein workingconditions.In mostEuropean countriesformal
technology
agreementsbetweentradeunions and employer organizations establish requirementsfor
developmentprocesses"
(Gr0nbaek, Grudin, B0dker,
andBannonp. 80).Therefore,
althoughPD worksinothercountries,theUnited States has adopted otherdesign
methods, whichlimitend userinput inthesoftwaredevelopmentprocess. "Although sustained userinvolvementseems
desirable,
itseffect on commercial productsisnot clear"(Grudin andPruitt 2002p. 1). This uncertainty abouttheimpactto thebottom line
limits theabilityof organizationstoembrace PD.
7. Scenario-Based Design
(SBD)
7.1
History
Theorigin of scenarios is in theatrical studies. Scenarios were and are used
by
manypeople such as militaryemployees,economistsandpolicymakers for
long
range planning andtoweigh theconsequences of actions (Carroll 2000ap. 320).However,
Ivar Jacobsen promulgatedthepopularityof scenarios withhis use caseapproach to
softwareengineering (Carroll andRosson 1992 p. 182).
In theearly
1980s,
SBD emergedfrom theHCIfieldtoresolve theneedtoprovidetherightbalance between modelingapproaches andpurelyexperimental
approachesthatrelied
heavily
on usabilitytesting
(Jarke,
Bui andCarroll 1998 p. 159).As HCIcontinuedto
develop
throughthe1980s,
focus was placed onthe concept of usabilityandits rolein systemdevelopment. Itwasbecoming
obvious thatusabilityofevery stepofthesoftwaredevelopment lifecycle startingwith analysis (Rossonand
Carroll 2002p. 13).
Inthe
1990s,
scenario-basedsoftware developmenttechniquesbecamepopularinsoftwareengineering because
they
describedkey
situationsin a narrativeformatthat anaudience representingvarious perspectives couldeasilyunderstand(RossonandCarroll
2002p. 7). Thiseliminatedtheneedto share complex models withtheend users and
stakeholders. Scenariotechniquesbridgethe communicationgap betweenthe team
membersthatprovide requirements andthe teammembers thatproduce andimplement
thedesign. Scenariosare a mechanismthatcanbeusedto streamlinetheexpectations
aboutthefinalproduct.
7.2 Definition
"Scenario-based designtries to
identify
the critical andtypical things thatpeople doorwanttodoandthe tradeoffsassociated withdesign featuresthatenablethese
activities"
(RossonandCarroll 2002p. 359). Scenariosfocuson thesituation
including
thepeopleinvolved,
the environment, theobjectives andthegoals. "Inscenario-baseddesign,
descriptionsofhowpeople accomplish tasks areprimary working design
representations"
(Carroll 2000bp.45). Scenario-based design containsfive
key
elements. Theinformation in Figure 2 definesanddescribestheseelements.
7.3 Techniqueand process summary
Thegoal ofthe SBDprocessis tofocusthedesign team's
thinking
on thenecessaryendusertasks (Rosson andCarroll 2002p. 315). The SBDprocessdoesthis
by
providinganiterativeapproachusing differenttypes of scenarios. Theprocessprovides theabilityto
refine theoriginally definedscenarios. The SBD processincludesthe
following
steps:(RossonandCarroll 2002p.
346)
1. Developmentof requirement scenarios
2. Validation/refinementof scenarios with users andcustomers
3. Developmentofbasic-leveltaskscenarios
4. Refinementofdesignscenarios withdevelopmentteammembersand
customers
5. Developmentofinformationmodel 6. Review withteammembers
7. Developmentof paper prototypes
10. Review withteammembers
11.
Develop
and refinearunningprototypes12. Formativeevaluation
13. Analysis oftranscripts/reportpreparation
14. Detaileddesignand prototype driven iterationof previous threesteps
4. Beprecise but include everyone on the team
Figure 2- Fiveelements ofScenarioBased Design
RossonandCarrollp.21
Thescenarios aredefined using a
five-step
method, whichis: analyzingtheorganizational
decision,
identifying
thekey
decisionfactors,
analyzingenvironmentalforces,
defining
thescenariologicandanalyzingtheimplications forthefinal decisions.Analyzeorganizational decisions isthefirst step and occurs priortocreating
scenarios. This is becausetheteammust understandtheorganization's goal and vision
forthefuturewithin thescope ofthe topic
being
analyzed. This analysisincludesinvestments,
and market strategies. Thesecondstepto thescenariobuilding
processis toidentify
thefactors thatinfluence an organization'sdecisions. These factorsprovideinsight intotheorganization'svalues thatcan
help
theprojectteamunderstandthestrategic goals and provide aframeworktomaketactical decisionswithin theproduct
developmentprocess. Thethirdstep isto analyze environmentalforces. This analysis
directly
defines future businessstrategy. This includes consideringpotentialcompetitorsand general economicconditions. Thebackgroundanalysis providesthemechanismto
createtheinitial scenarios. Thefourth step is todefinescenariologic. This step defines
thescenariotemplatethatwillbeusedthroughouttheprocess. "Scenariologic involves
organizing themes,principles, hypothesisand assumptions thatprovideeach scenario
with acoherent,consistent and plausiblelogic underpinning"
(Jarke,
BuiandCarroll1998p. 164). Thisthenmovestheprocess to thefifth andfinal step, whichis analyze
implicationsfor decisionsand strategies. This stepconsiders the implicationsofthe
scenario. The
following
questions shouldbeansweredinthis step:What dothescenariossayaboutthedesign and
timing
of particular strategies?
Whatthreatsand opportunities dothescenarios present
forthefuture?
Whatcriticalissuesemergefrom thescenarios?
Fromtheorganization'splanningperspective, what
kindof
flexibility
dothescenarios suggest arenecessary?
SBDprocessis hierarchal anditerative in its approach to
defining
andrefiningthedifferenttypesof scenarios. SBD defines fivescenario categories: problemscenarios,
activityscenarios,informationscenarios, interactive scenarios,and edge case scenarios.
Thescenario categories providetheanalysis focus forthe teammembers.
Aproblem scenariofocuses onlyon currenttasks. Theseare created with the
help
ofthestakeholders early in theprojectbecause
they
help
define thescope oftheproject(RossonandCarroll 2002p. 25).
Activity
scenarios (a.k.a.daily
usescenarios)focus ontheinformation theenduser needstocompletethetask. These scenarios attempttoresolve theissues defined in
workflow (RossonandCarroll 2002 p. 142). Thesescenarios definewhatneeds the
process shouldbewithout regardto thedesign.
Information scenarios are an extension ofactivity scenarios. Information
scenarios providethedetailabouttheinformationthe system providesto theend user.
Interactionscenarios describethe detailsoftheuser action andfeedback. More
explicitly interactionscenarios explain: theinformationneeds, theactionstheusertakes
tointeract withthe task
information,
andtheresponsesthe system providestothe user'saction. Theseare verysimilartousecasesonly ina more narrativeformat (Rossonand
Carroll 2002p.
16,
26).Thelasttypeofscenario, edgecasescenarios, shouldbeused cautiously. Edge
case scenariosincludetasks thatareuncommon, but analyzingthesecausesand effects
can prevent anincompleteproductdesign. In mostcases, thereisminimal valuein
creatingedge case scenariosbecause
they
have averynarrowfocus. As such,they
shouldonly be used whenthe time permits,orthelevelofcomplexity andimpactof
possible mistakes are socostlythat thesesituationscannotbe ignored (Cooper 1999p.
181).
Oncethescenarioshave been drafted
they
arefurtherrefined with theuse of claimsanalysis. Claimsanalysisis aformofparticipatory designwherethedesign teamand end
users scan the scenarioforthecauses andeffects andcapturethedesignrationaleforthe
future system. Claims analysisis defined
by
Chin, Rosson,
andCarroll(1997)
as "arefinementofparticipatory design in whichtheusers
directly
analyze current andenvisioned scenarios of
use"
(Carroll 2000bp. 272). Claimsanalysis provides a method
forsystematicquestioning thatis imperative inunderstandingtheuser's needs. This
techniquebringstolight
issues,
conflicts,andcontradictionsthatare notobviousinthescenario. Theproblem withclaims analysisis it is an iterativeprocess thatdoesnothave
a well defined startingorendingpoint(Carroll2000bp. 135). Thismeans thatit is
difficulttoknow whenenoughanalysis has beendone.
Typically,
thisactivity is timeboxedso thatwhenthe timeruns outthe analysisisconsidered complete.
The diagram in Figure 3 provides an overview of the scenario development
glance, itappears thatscenario developmentis alinearprocess, it is actually an iterative
process.
Scenario
Building
ProcessAnalyze
Analysisof
stakeholders, fieldstudies.
Metaphors, information
technology,HCI theory,guidlines
Summative
evaluation
Problem Scenarios
T
DesignActivityScenarios
InformationScenarios
InteractionScenarios
Prototypeand Evalute
Usabilityspecifications
Claimsabout current practice
Iterativeanalysis of
usablityclaims and redesign
Rosson& Carrolp.25
Figure3- Scenario
building
process1A Projectteam type
Cross-functionalteams workbestwhenusinganSBD process.
However,
becausetimelines areshrinking duetomarket
demand,
it isincreasingly
moredifficultto haveallteammembers participatein every stepoftheSBDprocess. Eachteammember
typically
when participatingon across-functionalteam.
Typically,
designteammembers areassignedtobuildcertain scenariosandthen the teamgatherstoshareknowledge.
7.5 Applicationdomain
ThetwoexamplesCarrollprovides suggests thatSBDis successfulon well
defined,
small-scaleprojects, using existing
technology
and acentrally locatedteams. Forexample,Carroll describesa project
delivering
a multimedia systemfor educatingengineers in a university. Theproject characteristics
include;
an enormousill-definedscope,
implementing
complextechnology, andgeographically distributedteammembers.Intheexample ofthemultimediaproject,Carroll explainsthata vision scenario was
createdbut only rarelyreferenced. Themultimedia systemproject veered offthe SBD
approach and used a prototypedrivenprocessbecauseoftheenormous scope. The
second exampleCarroll providesis adesignof a
library
systemthathadasmaller,well-definedscope, andcentrally locatedteammembers. Carrollexplainsthatthisprojectwas
also notcompletelysuccessful. The SBDprocess usedinthe
library
systemdesign didnotaccurately
identify
theroot cause oftheproblem andthis created rework. Carrolladmitsthat themultimedia and
library
examples were not representative ofbestcasepracticeforscenario-baseddesign (Carroll 2000bp. 43). Thiscallsintoquestionthe
capabilityofthismethod.
The capabilityofthis methodis furtherquestionedbecause Carrollexplainsthat
SBDhasnot used on projects thatincludesafety-criticalapplications, embeddedsystems,
orconcurrencyand,therefore, isnot sure ofits abilitytosupport suchprojects(Carroll
2000bp. 327). "Scenario-baseddesign may be
differentially
effectivefor different kindsofdevelopment
projects"
(Carroll 2000bp. 327). Carrolladmitsthat"scenariosare often
used
ineffectively
orinarticulately"
(Carroll 2000bp. 43). Although theremay be
benefitsofusing scenarios as atechnique,it issuspectthatSBDwill provide successful
results.
7.6 Integration
feasibility
Scenario-basedactivities canbe integrated into a system developmentmethodology.
"McGraw andHarbison
(1997)
have describedascenario-basedengineeringprocessthatintegrates scenariobaseddesign with traditionalstructureddevelopmentmethods, and
these traditionalmethods"
(Carroll 2000bp. 317).
They
explain thatscenarios closelymirrorhow software engineers andsoftwarequalityassuranceteammembers reason withintheirfunctionaltasks. Theclaims analysis will seem recognizabletoorganizations thathaveusedtraditional structured analysis methods tosoftware applicationdesign. This is becausesoftware claims are analogousto usabilityclaims usedinSBD. The
disparity
isthatsoftwareclaims addressimplementation featuresthathaveconsequenceson softwaredesignquality. The usabilityclaims articulatetherationalefor
designing
with end usersin mind.7.7 Advantages
Scenariosareeasytoproduceand assist with cross-functionalteamparticipation and informationsharing. Thenaturallanguageof scenarios allowsteammembersfrom
various perspectives toparticipate.
They
also provide a mechanismto shareknowledge within theteamandto transferknowledgetomembers outsidethe team.Scenarios
help
plan thework. "Thescenario-basedprocess allows alltheactivitiesto communicateina common vocabulary: orientinggoals andvisions,
workplace needs analysis, designwalkthroughs, envisionment, prototypes andrationales, evaluation of criticalincidents andtestcases, and potential generalizations and
abstractions"
(Carroll 2000b p. 309). Thereviewofthescenarios helpsthe team
members seethe project goalsin termsoftheend user's needs alongwith the
organization's goals. In addition,the scenariosfacilitatethecompletion of eachteam
member'sfunctionaltasks. Forexample, usage scenarios can
help
todefine interview questionnaires that theinteraction designer may berequiredtocreate.They
can alsobeusedto
help
guidetheimplementation anddeployment becausescenarios explainthefrequency
ofthe tasksand potential peaktimeswhenit maynotbeappropriatetoimplementa new system. Scenariosexpoundthepatterns andsocial mechanisms thatare
importanttoconsiderinthe design. Thepatterns oftaskcompletionin a scenario
identify
the appropriate
functionality
layout.Anotheradvantageisthatscenarios
help
designersreflectwhilegivingthemasense ofaccomplishment. "Ascenarioisaconcretedesignproposalthatadesignercan evaluate and
develop,
but is also roughinthatitcanbeeasily alteredandallowsmanytomovetowards amore concrete design. "Scenariosaregood atvividly capturingthe
consequences andtrade-offsofdesignsat variouslevelof analysisand withrespectto
various
perspective"
(Carroll 2000bp. 87). This leadsto thenextadvantage.
SBDprovides a systematic methodto designand evaluate. SBD facilitates design
rationaleusingclaimsanalysis. Theclaimsanalysis technique,whichis alarge aspect of
SBD,
makes therelationship between functionsand consequences of use explicit (Carroll2000bp. 152
-53). Claimsanalysis allows the team toanalyze explicitlythepros and
consof aspecific design asitrelatesto thescenario. Thesystematic method minimizes
downstreamdesignchanges.
SBD andits techniques matchthedesignanddevelopmentteam
members' cognitive
model making it easy forthem to learnthismethod. Carroll stipulates that
object-oriented software requires developersto thinkinterms offunctionsandSBD
complementsobject-orienteddesign reasoning because thevarioustypes of"scenarios
areideal structuresfor engineeringrequirementsandusability, that
is,
foridentifying
therequired
functionality
of systems andevaluatingtheeffectiveness oftheirrealization(Carroll 2000bp. 233).
7.8 Expectations
SBD techniqueshave beensuccessful and will continuetobe successful whenintegrated
into other user orienteddesign methods. Scenarios as atechniquehave beenborrowed
by
other user orienteddesign methodsbecause ofthevaluethis techniqueprovides to the
initial analysis phase.
However,
SBD as a user orienteddesignmethodhasnot yetproven itselfrobust enoughfor anytypeof project. Carroll admitsthat"it is far from
clearthatscenario-baseddesign will provide acompletelygeneral and comprehensive
systemdevelopment
methodology"
(Carroll 2000bp. 327). It isquestionableif SBDwill
everbeperceivedas a useable designapproachbecause itsrelianceon thescenario asits
main deliverable.
Oneofthechallenges for SBD isthattheprojectteammustbeableto tolerate the
vaguenessoftherequirements. Scenarios allow muchinterpretation ormisinterpretation
by
thereader. Softwareengineers customarilyhavealowtoleranceforambiguitywithinthe requirements, whichis themost probablereasonfor SBDnot
being
widely acceptedusedscenariosin thedevelopmentand analysis of prototypes
typically
aniterativeprocess ofrequirements
development."
(p. 317). Theuse of scenarios as an analysistool
works well;
however,
touse themastheprimary deliverabletoasoftware engineerhasnotbeen proven successful.
The informalnature ofSBD isyet another obstaclefor SBD. Carrollexplains,
"Scenarios are not
formal;
they
arenottechnicalin anyfancy
sense"
(Carroll 2000bp.
17). SBDtechniquesproduce words ratherthanpicturesor
diagrams,
andmosttechnicalteammembers donotenjoy readingvolumesofinformation. Thetechnical team
memberscannot see thepictureif
they
donot readthewords.In addition,therepeatablesuccess ofSBD hasnot yetbeen proven. Although
Carroll'sultimate goal istomakeSBD the
key
tomoving designworkintoascience,thisgoalisunrealistic. Theproblemis thatnotwosoftware developmentprojects areexactly
alike.
Lastly,
SBDsuccess dependsonthe training,support, and creation of scenariomanagementtools. There is a critical needfortools thatcreate, modify,organize and
track thescenario creation process. Scenarios canbecome difficulttomanage when there
aretoomany iterations and variations. Although scenariocreation onthesurface seems
easy,producingtheright number withtheright amount ofdetail is achallenge. In
addition,designerand software engineers needtolearn howto transform the scenario
intoa useful and usable softwaredesign.
8. Contextual Design
(CD)
8.1
History
Contextual Design
(CD)
begantoevolvein 1988. "CD wasdesignedatDigitalEquipment Corporation underthe
leadership
ofDr. KarenHoltzblatt"(Marion 1997a).
HoltzblattcreatedCD toprevent projectteamsfrom agonizingabout what shouldbe
designed
by involving
end users upon theprojectinitiation. In1992,
Karen HoltzblattandHugh Beyer founded InContext
Enterprise,
whichis aconsulting companythatisalignedwith theCD method. CDcontinues to
develop
asit isbeing
usedby
more8.2 Definition
ContextualDesign
(CD)
is "anapproachtodefining
software andhardware systemsthatcollect multiplecustomer-centeredtechniques intoan integrated designprocess(Beyer
1997p.3). The CD method providestechniquesfor gathering
data,
driving
design,
andmanagingteams. Thismethod stressestheneedtogather end userdata intheform of
datamodels thatbuildon a structural analysis approachusingtechniques such as object
modeling,enterprisemodeling, and process reengineering. Thetechniques
help
tofacilitatethedecisions abouthowtheend user willfunction in thefuture. "Contextual
Design provides completesupportforthedesignprocess,from theinitial customerdata
gatheringthrough the transitiontoobject-orienteddesignor whatever other
implementationmodel you favor"
(Beyer 1997 p. 5). Itprovides a clear processand
concrete actions. Thecentral aspect ofCD isthatitpromotes stayinggroundedinthe
understanding ofhowtheend user works.
8.3 Techniqueand processsummary
Thesteps for CDare as follows:
1. Contextual
inquiry
2. Work modeling
3. Consolidation
4. Workredesign
5. Userenvironmentdesign
(UED)
6.Mock-up
and evaluationThe initial step in CD is performingcontextual inquiries. Contextual
inquiry
is atechnique thatprovidesinsight intothecurrent workflow andincludes uncoveringtheend
user's tacitknowledge. Contextual
inquiry
typically
begins withone-on-oneinterviewsintheend user's workenvironment. An alternative approachtocontextual
inquiry
maybeneedediftheworktakes place overa
long
periodoris difficultto observe. Alternativeapproaches can be identified in Appendix A.
The subsequentstep in CD is creatingworkmodels. Workmodels consolidate
theinformationgathered
during
thecontextualinquiry
and provideatangiblerepresentationof end user's work. CDprovides severaldifferent modelingtechniques
thatassist theprojectteamsunderstandingthe
breadth
oftheend usersrequirements.By
moreflexiblesoftware applicationarchitecturethatsupportsfutureexpansion. In
addition, thesemodels areusefulin managingtheprojectandthemarketingofthe software application because
they
clearlydisplay
end user's workflowandthehigh-levelfunctionality. CDuses five typesof work models: workflowmodels, sequencemodels,
artifactmodels,cultural models and physical models.
Workflowmodels
help
teamstounderstandthe endusers'
communication
patterns. Theteamidentifiestherolesand responsibilities ofthepeople whodothework
andhow
they
interact. "Theconsolidatedflowmodelis thebest map tohow workisdone,
showingthebreadthof work andthe detailsofhow peopleinteract. Theflowmodel shows what roles peopleplay, andifyouhavesystems already inplace, youcan
see whatroles
[are]
supported"
(Beyer 1997p. 170).
Typically,
work rolesare consistentacrossdomainswhich can assisttheteamin providinga softwareapplication thatis
usefultoabroadrange of customersifthatis thegoaloftheproject.
Sequencemodels showtheorder of stepstaken tocompletethework,
including
the trigger thatinitiatedwork. "Sequencemodels show whatthecustomeris
trying
todoandhow
they
go aboutdoing
it"(Beyer 1997p. 197). Thesequence modelillustrates
communicationbreakdownswithin theprocess
being
analyzed. Thesemodels aresimilar to taskon arrowmodels anddecision diagrams thatare usedin structuredsoftwareanalysis approaches.
However,
thegoal ofthesequence modelsis notonlyto showtheorder ofthe stepstocompletethework,butalsotoillustratethecurrentcommunication
breakdowns.
Artifactmodelsinclude job aids,referencemanuals, andguides. Theartifacts
illustrate an opportunitytoenhancethecurrent software application. "These
[artifacts]
reveal howpeopleorganize and structure workfrom
day
today"(Beyer 1997 p. 178).Theartifacts are analyzedforstructure, usage, andintent. Thegoalofthisanalysisis to
integrate information intothe softwareapplication.
Culturemodels reveal work constraintscausedboth
by
anorganization'spolicyasa whole and values ofindividualemployees. The way employees dress andthelanguage
they
use defines an organization's culture. Thecultureisfurther definedby
therelationship between departmentsandthegeneralregardto thecustomers. "Thecultural models speakthewords peoplethinkbut don'tsay"
toemployees'
regardforeach other andthecustomer assiststhedesignteamto
understandtheorganization'sculture andprovides insightfor establishingthebasic
design standards. Forexample,ifa softwaretooliscreatedfora conservativecompany
thelookandfeel shouldnotinclude glitzygraphics.
Physical modelsrepresentthegroupingof people andequipment, andthe
movementbetweenobjects inordertogetthejobdone. Theenvironmentthatatoolis
usedin will often revealdesign constraints. Themovementofartifacts,communication,
toolplacement, amount ofspace, andtemperatureall affectdesign. Thismodel guides
thedesignteam'screation of asuitablemetaphor andtheplacementoffunctionswithin a
software application.
Eachmodelfocuseson a unique perspective to
help
the team stay focusedontheend
users'
issues. The flowmodelidentifiestheroles and responsibilitiesforsequence
model. Theresponsibilities arefurtheranalyzedto definethetasksthatmake upthe
sequences model. Artifactcollection identifiesthe toolsnecessarytocomplete one or
moreofthesteps inthe sequence. Thecombination ofthesemodelsis thenusedto
understandthecompanyculture, andthephysicallayoutoftheworkplace inrelationto
theend user's requirements.
CD providesa specific stepthatinvolves consolidatingthe
information,
making itmanageable enough tomoveinto adesign approach. Theconsolidation effort uses
interpretation sessions whereaffinity diagrams aredeveloped. The interpretationsession
includeparticipation oftheentireteamandbuilds a sense ofcommitmentinthe
developmentoftheappropriate solution. The interpretationsessions allow everyoneto
hearthesamequestions,answers, anddiscussions.
Therefore,
teammembersbuildoneach
others'
ideas. Beyerexplainsthat:
"Interpretation isthechain ofreasoningthat turns afact into an action relevantto thedesigner's intent. Fromthe
fact,
theobservableevent, thedesignermakes ahypothesis,
aninitial interpretation aboutwhatthefactmeans orthe intent behindthefact. Thishypothesis has animplication forthedesign,
which can berealized asaparticulardesign ideaforthesystem"
Therefore,
theinterpretation sessions provides theappropriateforumtodevelop
anaffinity diagramandconsolidatedwork models.
Affinity
diagramming
isone ofthe tools usedtoillustratetheconsolidated workmodels. The affinitydiagram
typically
hasahierarchical structureinwhich theaffinity isan outline oftheend user's work. It illustrates the
key
elementsoftheproblemthatthedesignteammust resolve. Theconsolidated work models
including
theaffinity diagramrepresentseeminglytrivialend user anecdotes as real system obstaclesthatrequire a
solution.
Theconsolidatedtasksequence model canbe donewiththeuse ofstoryboardsto
sketch outhow thesystem will worktoaccommodatethetask. Storyboards arerapidly
created and usedto illustratewith minimal detailhowthework willbe done. This
techniqueprovidesthe teama visual aidforwhathastohappen inthedesign and whenit
hastohappen.
Workredesignis thenext appropriate step if it iswithinthescope oftheproject.
Workredesign provides efficiencies to thework practicethatmayormaynotincludethe
use oftechnology. This stepalso takes advantage ofthestoryboardtechniquebecause it
visuallycommunicatestheworkflow.
Oncethework models areconsolidated, theuser environmentdesign
(UED)
model is created. CDuses aUEDmodeltoillustrate howeach part ofthesystem
supportstheend user'swork,whatfunctions areavailable, andhow
they
areinterconnected. The UED isnotconcernedaboutthe UI
details,
but ratherthefunctionsthat thesystem will provide andhowthesefunctionsrelatetoeach other. Thefunctions
arethen made availablethrough menus,toolbars, andkeyboardcommands thatcan be
consideredlater inthe designprocess. Ifstoryboards werecreated
they
can be leveragedto buildtheUEDmodel.
However,
theUEDmodel can be builtwithoutstoryboardsifacurrentsystemis inplaceandtheprojectgoalis toreengineerthecurrentsystem. Inthis
situationthe UED isusedtorevealthecurrent system's problems in meetingtheend
user's workflowrequirements. Similarto the storyboard, the UED isusedtomanagethe
structureofthe system; thedifferenceis theUED adds alevelofdetail.
The UED canbeusedforseveral purposes
depending
onthegoal oftheproject.release. This is becauseit illustrates thescope of
functionality
fortheend-to-end workprocesses
being
analyzed. Whenthegoalistocreate a new softwareapplication,theUEDprovidestheteamaconcrete and consistentunderstandingof whatdataneedsto be
available tocompleteeach activitywithin a given process. The UED specificallyassists
theinteraction designers in creatinganappropriatelylayeredinteractiondesign because
"the UED works against proliferation of
dialog
boxes"
(Beyer 1997p.398). IftheUED
has allottedforafunctiontobe part ofthefocusarea, then thatfunction musthave
allocated spacein thedesign. The UEDcan alsobeusedtoleve