11
1 Robertson, T ‘Cooperative Work and Lived Cognition: A Taxonomy of Embodied Actions’ in J A Hughes (ed.) Proceedings of ECSCW’9, Dordrecht, Kluwer (1997) pp 205-220. 2 Dourish, P Where the action is MIT Press, Cambridge, MA (2001) 3 Tang, J C ‘Findings from observational studies of collabor- ative work’ International Journal of Man Machine Studies Vol 34 No 2 (1991) 143–160 4 Radcliffe, D F ‘Ideas and knowledge displays within a design team’ in N Cross, H Chri- stiaans and K Dorst (eds) Ana- lysing Design Activity, Wiley, Chichester (1996) pp 343–364 5 Olson, G M, Olson, J S, Carter, M R and Storrøsten, M ‘Small group design meetings: an analysis of collaboration’ Human–Computer Interaction Vol 7 (1992) 347–374 www.elsevier.com/locate/destud 0142-694X $ - see front matter Design Studies 24 (2003) 537–547 doi:10.1016/S0142-694X(03)00041-3 537 2003 Elsevier Ltd All rights reserved Printed in Great Britain Telling tales: understanding the role of narrative in the design of taxonomic software Susan Turner and Phil Turner, School of Computing, Napier University, 10 Colinton Road, Edinburgh EH10 5DT, UK This paper draws on concepts from narratology to explore aspects of the role of stories in the small group design process. A brief review of relevant narratological concepts is provided. Their application in an analysis of case study data from a team designing taxonomic software is then reported. It is concluded that narratology, and in particular the notions of focalisation and the active listener, has useful descriptive potential in this context, and may help to elucidate some difficulties in design communication and documentation. Suggestions for extension of the work are included. c 2003 Elsevier Ltd. All rights reserved. Keywords: case study, communication, design process, narrative, protocol analysis I n this paper we present a case-study based exploration of the role of stories in the small group design process. Small group design activity has been the subject of much research and theoretical development. Shared space, physical resources and embodied action (e.g. the work of Robertson 1 , Dourish 2 , Tang 3 and Radcliffe 4 ) have received considerable attention, as have the study of the elements and dynamics of the design process (as in Olson et al. 5,6 ; Potts and Catledge 7 , Ball and Ormerod 8 ) while the social constructivists have also made their case (e.g. Bucciarelli 9 and Lloyd 10 ). Again from a social perspective, collaboration styles have been studied by Maher et al. 11 while the pattern, structure and substance of communication within and outside the team has received the attention of a number of researchers (e.g. Cross and Clayburn-Cross 12 , Cahour and Pemberton 13 , Perry and Sanderson 14 , Reid and Reed 15 ). We contribute to this last theme by illustrating how stories steer the design work, act as communication tools and create obstacles to mutual understanding. This last point is explored and illustrated by way of a narratological analysis. The paper shows how such an analysis can help both to describe and under-

Telling tales: understanding the role of narrative in the design of taxonomic software

Embed Size (px)

Citation preview

Page 1: Telling tales: understanding the role of narrative in the design of taxonomic software

1 Robertson, T ‘CooperativeWork and Lived Cognition: ATaxonomy of Embodied Actions’in J A Hughes (ed.) Proceedingsof ECSCW’9, Dordrecht, Kluwer(1997) pp 205-220.2 Dourish, P Where the actionis MIT Press, Cambridge, MA(2001)3 Tang, J C ‘Findings fromobservational studies of collabor-ative work’ International Journalof Man Machine Studies Vol 34No 2 (1991) 143–1604 Radcliffe, D F ‘Ideas andknowledge displays within adesign team’ in N Cross, H Chri-stiaans and K Dorst (eds) Ana-lysing Design Activity, Wiley,Chichester (1996) pp 343–3645 Olson, G M, Olson, J S,Carter, M R and Storrøsten, M‘Small group design meetings:an analysis of collaboration’Human–Computer InteractionVol 7 (1992) 347–374

www.elsevier.com/locate/destud0142-694X $ - see front matter Design Studies 24 (2003) 537–547doi:10.1016/S0142-694X(03)00041-3 537 2003 Elsevier Ltd All rights reserved Printed in Great Britain

Telling tales: understanding the roleof narrative in the design oftaxonomic software

Susan Turner and Phil Turner, School of Computing, Napier University,10 Colinton Road, Edinburgh EH10 5DT, UK

This paper draws on concepts from narratology to explore aspects ofthe role of stories in the small group design process. A brief review ofrelevant narratological concepts is provided. Their application in ananalysis of case study data from a team designing taxonomic software isthen reported. It is concluded that narratology, and in particular thenotions of focalisation and the active listener, has useful descriptivepotential in this context, and may help to elucidate some difficulties indesign communication and documentation. Suggestions for extension ofthe work are included.�c 2003 Elsevier Ltd. All rights reserved.

Keywords: case study, communication, design process, narrative,protocol analysis

In this paper we present a case-study based exploration of the role ofstories in the small group design process. Small group design activityhas been the subject of much research and theoretical development.

Shared space, physical resources and embodied action (e.g. the work ofRobertson1, Dourish2, Tang3 and Radcliffe4) have received considerableattention, as have the study of the elements and dynamics of the designprocess (as in Olson et al.5,6; Potts and Catledge7, Ball and Ormerod8)while the social constructivists have also made their case (e.g. Bucciarelli9

and Lloyd10). Again from a social perspective, collaboration styles havebeen studied by Maher et al.11 while the pattern, structure and substanceof communication within and outside the team has received the attentionof a number of researchers (e.g. Cross and Clayburn-Cross12, Cahour andPemberton13, Perry and Sanderson14, Reid and Reed15). We contribute tothis last theme by illustrating how stories steer the design work, act ascommunication tools and create obstacles to mutual understanding. Thislast point is explored and illustrated by way of a narratological analysis.The paper shows how such an analysis can help both to describe and under-

Page 2: Telling tales: understanding the role of narrative in the design of taxonomic software

6 Olson, G M, Olson, J S,Storrøsten, M, Carter, M,Herbsleb, J and Rueter, H ‘Thestructure of activity during designmeetings’ in T P Moran and J MCarroll (eds) Design Rationale:Concepts, Techniques and Use,Lawrence Erlbaum, Hillsdale, NJ(1996) pp 217–2397 Potts, C and Catledge, L‘Collaborative conceptualdesign: a large software projectcase study’ Computer SupportedCooperative Work Vol 5 No 4(1996) 415–4458 Ball, L J and Ormerod, T C‘The influence of co-designerson the generation and evaluationof solution alternatives’ in S A RScrivener, L J Ball and AWoodcock (eds) CollaborativeDesign, Springer, Dordrecht(2000) pp 243–2529 Bucciarelli, L L DesigningEngineers MIT Press, Cam-bridge, MA (1994)10 Lloyd, P A ‘Storytelling andthe development of discourse inthe engineering design process’Design Studies Vol 21 No 4(2000) 357–37311 Maher M L Cicognani Aand Simoff S J ‘An experimentalstudy of computer mediated col-laborative design’ InternationalJournal of Design Computing(1997)12 Cross, N and Clayburn-Cross, A ‘Observations of team-work and social processes indesign’ in N Cross, H Chris-tiaans and K Dorst (eds) Ana-lysing Design Activity, Wiley,Chichester (1996) pp 271–29913 Cahour, B and Pemberton,L ‘When people design: conver-sational positioning and roles incollaborative design dialogues’in F Darses and P Zarate (eds)Proceedings of COOP’98,INRIA, Sophia Antipolis (1998)pp 35–4614 Perry, M and Sanderson,D ‘Co-ordinating joint designwork: the role of communicationand artefacts’ Design StudiesVol 19 No 3 (1998) 273–28815 Reid, F J M and Reed, S E‘Interaction and entrainment incollaborative design meetings’ inS A R Scrivener, L J Ball andA Woodcock (eds) Collabor-ative Design, Springer, Dord-recht (2000) pp 233–24216 Turner P Turner S andMcCall R ‘Getting the storystraight’ in People and Com-puters XV, Proceedings ofHCI’01 (2001) pp 267–278

538 Design Studies Vol 24 No. 6 November 2003

stand the process and the communication difficulties which arise. Ouranalysis (and illustration) is based on a study of a small group of designersdeveloping taxonomic software for botanists in a project known as Prome-theus II.

As practitioners and occasional theorists of user-centred design (UCD),this paper forms part of a developing strand of research into the role ofnarrative in the user-centred design process. In contrast to the more familiarstudy of narratives in design (i.e. the use of the ubiquitous scenario) weare interested in how the concepts, vocabulary and constructs of narratol-ogy can provide us with a means of analysing, describing and illustratingthe subject matter and dynamics of design meetings. Elsewhere we havediscussed how the desire to commit to a coherent joint design story sub-merges many competing narratives, with predictably unfortunate results16.We have also shown how a strong narrative structure is an essential compo-nent of a successfully engaging collaborative virtual environment17 andhow the actions of the designer as narrated in many reports of UCD bearsa startlingly close resemblance to the stylised story of the hero’s journeyfound in myth and folktale18.

Our focus complements the work of Lloyd10, who identifies four main rolesfor stories in the design context: in giving an account of progress to others;as a convenient index to past events or discussions, as competing rationalesfor the same object—so a specification may be regarded as a flexibleresource for efficient selling in the salesman’s discourse, but a definedstarting point for the designers—and as the socially constructed meta-storyof the process itself. We are specifically concerned with the role of stories(or story fragments) which are told by designers about actual and potentialuser behaviour and about the role of the software under development. Theanalysis draws on structural concepts from narratology to analyse how thedesign team tell stories which arise from the same set of agreed userrequirements, but are different in subtle, often unacknowledged, but insidi-ous details.

1 A very brief introduction to narratologyOf necessity, this paper can only touch upon the richness of narratologicaltheory, and the reader is referred to publications such as Onega andLanda19, Bal20 and Herman21, for thorough introductions. For our purposes,we use narratology to mean the study of narrative texts, the ‘…represen-tation of a series of events meaningfully connected in a temporal and causalway’19. A ‘text’ may be in any medium, thus encompassing the spokenword, film and pictures as well as written material. In recent decades narra-tological analysis has been adopted in many disciplines beyond its original

Page 3: Telling tales: understanding the role of narrative in the design of taxonomic software

17 Turner P and Turner S ‘Anaffordance-based framework forCVE evaluation’ in People andComputers XVI, Proceedings ofHCI’02 (2002) pp 89–10318 Turner S and Turner P‘Designer as Hero’ in ‘Voicesfrom the Fringe’ workshop,HCI’01 (2001)19 Onega, S and Landa, J SG (eds) Narratology, Longman,London (1996)20 Bal, M Narratology: intro-duction to the theory of narrative,2nd edn University of TorontoPress, Toronto (1997)21 Herman, D (ed.) Narratolog-ies :New Perspectives on Narra-tive Analysis, Ohio State Univer-sity Press, Columbus, OH (1997)22 Czarniawska, B Narratingthe Organization: Dramas ofInstitutional Identity University ofChicago Press, Chicago, IL(1997)23 Pentland, B ‘Building pro-cess theory with narrative: fromdescription to explanation’ Acad-emy of Management Review Vol24 No 4 (1999) 711–724

539Narrative in taxonomic software

domain of literary theory, most pertinently for us in discussions of thedesign process10 (e.g. Lloyd ibid) and organisational theory22,23. Our treat-ment draws on the structuralist tradition of narratological studies. Post-structural theorists have addressed, inter alia, the central but vexed ques-tion of the relationship between author and reader in the construction of astory, but this is beyond the scope of the preliminary analysis presentedin this paper.

1.1 Some basic conceptsWe consider design communication in relation to the three levels of narra-tive commonly adopted by literary narratologists, at least those of a struc-turalist persuasion. To paraphrase Bal’s definition20, these are:

� the text itself—the set of linguistic signs which appear on the page, orare spoken or otherwise presented;

� the story which the text conveys;� the fabula—the underlying events and circumstances of which a given

story is only one possible account.

In our case we will consider the taxonomical software project outline asthe fabula, and the story level as containing the set of explanations of howtaxonomy is currently practised, how the new software will support (orchange) this process, how the software will be used, how the software willwork and so forth. The text consists of the words spoken by the designers,recorded as the verbal protocol of the meeting, together with the documentsauthored by project members (not considered further in the current paper.Our discussion very largely concerns an exploration of the story level inthe design meetings studied.

The structural analysis of stories in literary narratology has focused on awide range of story features, including but not limited to: characters, theircharacteristics and values; different types of narrator and the relationshipbetween narrator and point of view; the actions available to particular typesof character; the cross-cultural persistence of a set of story themes; thereflection of cultural norms in what is valued, condemned or left unsaid;the role of the narratee (the reader, viewer or listener) in shaping the story;intertexuality (the way in which all stories necessarily relate to all otherstories known to author or reader); and the treatment of temporal perspec-tives. We focus here on two main aspects: the differing emphases in thestories told by designers, and the manner in which these stories are toldand heard from different perspectives. A little more should be said hereabout the treatment of perspective in narratology.

Page 4: Telling tales: understanding the role of narrative in the design of taxonomic software

24 Lodge D ‘Analysis andinterpretation of the realist text’Poetics Today (1980) 1.425 Moro, Y ‘The expandeddialogical sphere: writing activityand authoring of self in Japaneseclassrooms’ in Y Engestrom, RMiettinen and R -L Punamaki(eds) Perspectives in ActivityTheory, Cambridge UniversityPress, Cambridge (1999) pp165–182

540 Design Studies Vol 24 No. 6 November 2003

1.2 Focalization, utterances and voiceEvery story has a narrator. This may be external, whether the author of anovel or an academic paper, the witness taking the stand in the courtroom,or the teller of an anecdote in the coffee room. Equally, the narrator maythemselves be a character in the story. However, the story is necessarilytold from someone’s (occasionally something’s) perspective, and this isnot always that of the narrator. It is the owner of this perspective whomBal20 and other theorists term the focalisor. Sometimes the change offocalisor is explicitly introduced (e.g. ‘I asked her to tell me how it allstarted and she said…’), but frequently the adoption of a focalising personais subtle and un-signposted. Just as for the characters in the story, eachfocalisor has their own set of characteristics and values. Events are thusseen and told through the a filter of perceptions and conceptions24 belong-ing to the focalisor, as well as the narratological lens.

Focalisation can be regarded as a variant on the work of the Russian aca-demic, Mikhail Bakhtin (1895–1975) on the role of ‘voice’ and the relatedconcept of utterance. Bakhtin argued that speech comprises utteranceswhich are always oriented towards a particular addressee. An utterance isspeech terminated by a change of speaker, and reflects a point-of-view or‘voice’ in Bakhtin’s terminology. Thus in the context of user centreddesign, language mediates the interplay between designers and designersassuming the role of users with the utterance being the appropriate unitof analysis.

A key element of Bakhtin’s argument is that both speaker and listener areactive in the process of constructing meaning—there are resonances herewith post-modern narratology. Bakhtin observed that each speaker (or,author of an utterance as he phrased it) constructs each utterance as aresponse to ‘other viewpoints, world views, trends, theories, and so forth’embodied in the utterance to which he or she is replying. This turn takingbetween speaker and listener is active on both sides: ‘the person or personsfrom whom the author (speaker) expects a response—‘the addressee’—isan active participant for whom the entire utterance is constructed in antici-pation of the response’ Thus utterances are ‘determined by whose word(utterance) it is and for whom it is meant’ (quoted in Moro25, p 86.)

The notion of an utterance is closely linked to that of a voice—it reflectsa point-of-view which is Bakhtin’s definition of a voice. A Baktinian voiceis similar to the English use of the word ‘hat’ to indicate a temporarilyadopted role or point-of-view—as in ‘speaking with my research hat on Ithink x, but with my teaching hat on I believe y’. Voices exist in a socialmilieu and cannot exist outwith a social context. Meaning can only come

Page 5: Telling tales: understanding the role of narrative in the design of taxonomic software

541Narrative in taxonomic software

into existence when two or more voices interact—the voice of the speakerand the voice of the listener. Thus Bakhtin insisted that both voices shouldbe taken into account.

Having equipped the reader with a set of narratological concepts andvocabulary we now turn to our case study.

2 The Prometheus II case studyThe purpose of the Prometheus II project is the design of a data modeland a database system that would support understanding and comparisonof plant descriptions by taxonomists. Currently, taxonomic descriptions donot use words and concepts consistently. For example multiple definitionsof leaf exist, varying with the taxonomist and the plant group described.Clearly, this prevents unambiguous communication impossible and thusfosters misunderstanding. The Prometheus approach comprises a standardstructure for descriptions, so their processing by a computer system is poss-ible. All possible descriptions can be captured in the structure. In addition,terms are always accompanied by a definition, so that taxonomists canaccess the definition of a term to make sure their understanding correspondswith the intentions of the author of the description. This also supports thediscovery of implicit relationships by reasoning about terms and conceptsand more accurate and valid comparisons.

2.1 Dramatis personaeAt the time of the first meeting in the series studied (see next section) theproject was just starting to design the first in a series of three conceptualdata models for plant classification. The model was to be used to drive thedesign of an initial software prototype for testing with users and with aspecimen data set. Further iterations would develop two further data mod-els and three prototypes of increasing degrees of complexity and sophisti-cation. The core project members taking part in all the observed meet-ings were:

� E—the project manager at institution A (a school of computing), a data-base specialist, although originally a biologist

� F—the project manager at institution B (a botanical institute), a botanistand taxonomical software specialist

� G—a research assistant at institution A, a database specialist� H—a research assistant at institution B, a biologist.

Other personnel contributed to some meetings, specifically another data-base specialist and a research student investigating data visualisation issues,both members of institution A.

Page 6: Telling tales: understanding the role of narrative in the design of taxonomic software

542 Design Studies Vol 24 No. 6 November 2003

3 The approach to analysisWe should stress here that this was a preliminary study to explore whetherand how narratological concepts could be applied to inter-team communi-cations in the design phase. One of the first two authors attended each offive successive meetings between designers in the project. Meeting 1 wasaudio-taped, and meetings 2 and 3 videotaped. Problems with equipmentprevented recording of meetings 4 and 5, but the attending author tookdetailed written notes, including verbatim records in instances where storiesor fragments of stories were voiced. Meetings varied in length, but in totalmaterial from around 7 hours of meetings was collected. Informal question-ing of the second two authors (members of the design team) by the firsttwo authors, together with project documentation, provided the context forthe data obtained from meetings.

The approach to subsequent analysis (by the first two authors) was bothbottom-up and top-down. The meeting data (for this study confined to thecombined verbal protocol generated by the participants, forming the textin narratological terms) was reviewed several times, at first to gain famili-arity with the content, and on later passes to make a preliminary identifi-cation of aspects of storytelling in the meetings. Once stories or story frag-ments had been identified, a more fine-grained narratological analysis wasapplied using the concepts from the literature identified above. The nextsection summarises the preliminary results of this work, drawing out thevarying roles of story in this particular instance of design process.

4 ResultsEach of the meetings comprised discussion (or continued discussion) of aseries of draft documents prepared by G, the computing research assistant.The documents were intended to capture the current understanding of theconceptual data model for classifying plant specimens. In later months, theproject would turn its attention specifically to usage aspects, but even atthis stage there was much discussion of the concrete details of current andpotential taxonomic practice. Our first step was to interpret and form anunderstanding of the text we had captured. This process is necessarily andunavoidably subjective and the end result is a series of different storiesabout the design process. This analysis of the meeting data uncovers evi-dence of differing stories on several levels. These are now discussed, withverbatim extracts from the meeting protocol (in italics) included by wayof illustration.

4.1 Variations on a themeThe analysis of the text (the collective verbal protocol) reveals there areat least two stories concerning the purpose and direction of the project in

Page 7: Telling tales: understanding the role of narrative in the design of taxonomic software

26 Napier University Prome-theus II: capturing and relatingcharacter concepts in plant tax-onomy (2002) http://www.dcs.napier.ac.uk/~prometheus/prometheus 2/prometheusII.html

543Narrative in taxonomic software

terms of the benefits to be gained from the taxonomic software, and towhom these benefits will accrue. These do not conflict fundamentally, andindeed are both present in one of the high level project descriptions26,but their differing emphases have consequences when the practical usageimplications are considered. These parallel variations distilled from themeeting protocols can be summarised as set out below.

Variation a—The taxonomist as expert. The taxonomic classification of botanical

specimens is a complex task, requiring the exercise of expert judgement. The software

will support taxonomists in this task and be as easy to use as possible.

Or as the project description26 states ‘…one of the main challenges in theproject will be to create a user-interface that allows the specification andsubsequent use of the character framework in an easy way for taxonomiststo use. Statements regarding the equivalence of concepts must remainunder the control of the taxonomist…’

And in parallel

Variation b—The need for enforcing consistency. Classification of specimens in

current taxonomic data is inconsistent; this makes comparisons and reuse of

descriptions difficult. The software will help the taxonomic community by enforcing

consistency.

Or again from the project description ‘..a three year project concernedwith increasing the rigour and precision with which taxonomists can store,integrate and disseminate their data… …taxonomic output is fundamentalto all fields of biology that refer to organisms, and taxonomists in turn usedata derived from these sources when refining past classifications. How-ever, in taxonomy there is no single way of interpreting data, even thoughrigorous scientific methods are used to collect them. The result is confusionwhen using names and ambiguity when reading descriptions…’

The above variations are rarely voiced explicitly during the meetings, butsurface in the designers’ discussions of how the software will be used.Bakhtin’s concept of the active listener proved useful in analysing the tran-scripts for underlying stories. Not only are do the stories belong to differentfocalisors or voices, but the stories and their component utterances areaddressed to different listeners. The listeners to the utterance below—partof the ‘consistency story’—seem to be the designers in character as mem-bers of the Prometheus project, rather than as proxy taxonomists, or data-base designers, or another of the roles discussed in the next section.

G. …we can choose to trust users, but do we want to, that’s my only question.

As we shall see in the next section, the different stories generally belong

Page 8: Telling tales: understanding the role of narrative in the design of taxonomic software

544 Design Studies Vol 24 No. 6 November 2003

to different focalisors as they described current practice and potentialsoftware use.

4.2 Characterisation and focalisation in usage storyfragmentsContributions to discussions about current and future taxonomic practiceare frequently expressed as fragments of stories. The fragments concernthe interplay between several characters: the taxonomist, both as a prac-titioner using current tools and as a future user of the software; the softwareitself; and a shadowy entity which we might characterise as the ‘taxonomiccommunity. From a close analysis of the story fragments we can see thecharacteristics ascribed to each character. For example, taxonomists arevariously idiosyncratic, trustworthy, rigorous, and require flexibility. Hereis an illustrative extract from meeting 1.

F: …so the only sensible way forward is to give the user some credit and hope that

they’re going to put appropriate units.

Taking the analysis a stage further, stories of taxonomists and the softwarecan be seen to adopt different focalisors. (As we briefly explained above,a focalisor is the character from whose perspective events are perceived.)All the four designers frequently narrate fragments about how taxonomistscurrently work, or how they will do so once the software is in place. Inthe meeting protocol, the designers most frequently focalise their remarksfrom their own primary perspective, whether this is as taxonomist or data-base designer. This is H, a biologist, describing the practice of measure-ment in relation to landmarks in meeting 3. She is placing herself in therole of the taxonomist recording the measurement

H: Landmarks are used to highlight where you’ve made a measurement …

And E, a database specialist, commenting about taxonomic practice but onthe evidence of the terminology used, speaking from her own perspective…

E: So the taxonomists, when they’re describing something, they’re entering structure,

property, value…

However, the database specialists sometimes focalise the perspective ofthe prospective user of the software (extract below), and the biologists inturn speculate about how the software might work from a technical per-spective.

G: When you read the description at breast height you don’t have a picture of the

person who described it so how do you know how high it is…

Page 9: Telling tales: understanding the role of narrative in the design of taxonomic software

27 Bartlett, F C Remembering:an experimental and social studyCambridge University Press,Cambridge (1932)

545Narrative in taxonomic software

And quite frequently the identity of the current focalisor is unclear—who,for example are ‘we’ in the fragment below?

G: and we have problems for comparisons because we have to understand what

breast height means…

The values implicit in different variations of the project story, and thecharacteristics ascribed to users surface in how users actions are described.When the focalisor is a user, flexibility and ease of use tend to be empha-sised, even if the speaker is a database designer, as in the second extract.(The extracts are discontinuous.)

H: You have to put a unit on, but you are allowed to choose what unit you put on

there.

E: …they don’t stop and think am I doing one like this or doing one like this or

doing one like this…

Conversely when speaking as a database designer or member of the taxo-nomic community, expressed values are of consistency and clear structure,while users are characterised as idiosyncratic, as in the statement belowfrom one of the biologists. Note that this is also a very rare instance of‘point of view’ being mentioned explicitly.

F: I think you see from point of view of reproducibility and consistency then if you

just let them use free text then you start to lose meaning and comparability.

Thus we have a situation in which parallel versions of stories of purposeand use, with their attendant values, and addressed to different listeners,flow through the meetings side-by-side. In the meeting itself, we suggestthat designers are aware of these subtle differences, and although navi-gation through the eddying streams and currents of story can be a lengthyprocess incorporating diversions and false starts, there is an impression ofconsensus. However, this fragile achievement has to be created anew atthe beginning of each session. At least two factors can be identified atwork here. Firstly, between most meetings discussion is filtered and con-densed onto paper, a medium much less forgiving of parallelism, ambiguityand inconsistency then ephemeral speech. Secondly, a further problemhampering the achievement of mutual understanding in the design meetingsis the nature of our cognition. As early as 1932, Bartlett27 demonstratedexperimentally that stories are reconstructed rather than recalled. When heasked people to read and remember a story he found that of the originalmaterial little was retained except for ‘isolated and striking details’. More-over these details were only remembered if they fitted into the individual’spre-conceptions or prior experience. In practice this means that memoriesare recalled as gist.

Page 10: Telling tales: understanding the role of narrative in the design of taxonomic software

28 Conklin, E J and Burgess-Yakemovic, K C ‘A process-ori-ented approach to design ration-ale’ Human–Computer Interac-tion Vol 6 (1991) 357–39229 MacLean, A, Young, R M,Bellotti, V M E and Moran, T P‘Questions, options and criteria:elements of design space analy-sis’ Human–Computer Interac-tion Vol 6 (1991) 201–250

546 Design Studies Vol 24 No. 6 November 2003

5 DiscussionIt is noticeable from reviewing the meetings that discussions of apparentlyquite small points of envisaged usage, and the conceptual model that suchusage would entail, are very prolonged. This is confirmed by one of thedesigners, who noted that the first version of the conceptual data modelhad taken rather longer to agree than originally envisaged. Any causalrelationship between different perspectives adopted by designers, the dif-fering values that belong to those perspectives, and the different contextsof story fragments, and the elusiveness of clear resolutions cannot of coursebe proved. But we suspect at least some obscurity was contributed by thesenarratological phenomena. In particular, the concept of focalisation at thelevel of usage stories has allowed us to suggest some of the reasons forprolonged discussion. (We should add here that we do not believe that theproject studied was exceptional in its deliberations. Most of the manydesign projects with which we have been involved have encountered simi-lar difficulties.)

The subtle and multi-stranded nature of design narrative we have illustratedabove adds an additional explanation for the lack of systems such asgIBIS28 and QOC29 which attempt the capture of formal and unambiguousdesign rationale. The central role of the story in the design process alsosupports the growing tendency within the human computer interactioncommunity to document designs by scenario. There would appear to bescope for a suite of parallel scenarios, embodying stories told by differentfocalisors and with different emphases, but which nonetheless are satisfiedby the eventual design solution.

As for the next steps in this work, we intend to extend the preliminaryanalysis described in this paper by (i) examining the continuity of storiesbetween meetings and documents (ii) tracking any systematic change infocalisation of usage stories over time, and (iii) by applying the conceptsof the active listener or narratee to the data. We will then refine the analyti-cal concepts for application to a more extended and complete series ofdesign meetings. In the longer term, there are other aspects of story toaddress. In particular, we suspect that that there may be stereotypical stories(genres) of users and usage which cross different design domains.

6 Afterword: the academic storyWithin the text of the meetings, there are explicit references to the projectstory that to be presented to the outside world, more specifically the com-munity of fellow researchers. Here the issue is one of deciding the empha-sis of a formal story for publication in academic forums. For example hereare E and F discussing the academic story in meeting 2.

Page 11: Telling tales: understanding the role of narrative in the design of taxonomic software

547Narrative in taxonomic software

E: As far as I’m concerned the important conceptual idea is that we’re tracking the

usage of defined items, [murmurs of agreement] that’s what we need to make sure is

captured in the conceptual model and published...

F: That’s my view as well.

This is not just a matter of presentation or ‘spin’; the plot of the agreedacademic story will partially determine the focus of the project’s work.Thus, just as in the writing of a novel, the project’s fabula (or activities)will be arranged to match an agreed story.

AcknowledgementsWe are indebted to Jessie Kennedy and Cedric Raguenaud for critical dis-cussion of the ideas in an earlier version of this paper. We are also gratefulto them and their Prometheus II colleagues for allowing us to use datafrom their meetings.