rapid knowledge construction: a case study in corporate contingency planning using collaborative...

& Case Study

Rapid Knowledge Construction: A CaseStudy in Corporate ContingencyPlanning Using CollaborativeHypermedia

Albert M. Selvin1* and Simon J. Buckingham Shum2

1Verizon Communications, White Plains, NY, USA2Knowledge Media Institute, The Open University, UK

Many knowledge management (KM) efforts revolve around managing documents in arepository or enabling better real-time communication. An ideal approach would combinethese with the ability to create knowledge content that can be either formal or informal in nature,in a rapid, real-time manner. We will call this Rapid Knowledge Construction (RKC). This paperdescribes the concepts underpinning our approach to RKC, and provides a case study of theapproach in an industry context. The Compendium approach, which has been applied in projectsin both industry and academic settings, facilitates the rapid creation of the content of a KMrepository, by combining collaborative hypermedia, group facilitation techniques, and ananalytical methodology rooted in knowledge acquisition and structured analysis. Compendiumaddresses key challenges for the successful introduction of KM technologies into work practice:(1) customization for different use contexts; (2) integration of formal and informal communica-tion; (3) integration of both prescribed and ad hoc representations; (4) validation and cross-referencing of the repository ‘on the fly’ at the point of entry; (5) conversion of organizationaldocuments/emails into a hypertext database, and (6) conversion of hypertext databases intoorganizational document formats. Copyright # 2002 John Wiley & Sons, Ltd.

INTRODUCTION

The Knowledge Management (KM) field is satu-rated with digital technology, with every concei-vable system being rebadged and marketed as a‘KM solution’. However, while we all recognizethe marketing pitch of hardware and softwarevendors, given the breadth and fuzziness of theKM concept we should not be surprised if almostall technologies can carve a niche somewhere inthe ‘knowledge chain’ from raw data to mean-ingful knowledge resource. The question of KM’ssignificance as a new field is an ongoing dialogue.

This paper’s contribution to that debate is twofold:we draw attention to a number of concepts whichwe find help to sharpen the idea of knowledge-leveltechnologies, as distinct from the traditional con-cerns of data processing and information systems;second, we show their relevance to a field-testedmethodology and software tool. We begin with theconcept of Rapid Knowledge Construction, fol-lowed by the key perspectives that motivate theapproach and case study that we present.

RAPID KNOWLEDGE CONSTRUCTION

Many knowledge management efforts revolvearound managing documents in a repository or

*Correspondence to: Albert M. Selvin, Verizon Communications,White Plains, NY10604, USA. E-mail: [email protected]

Knowledge and Process Management Volume 9 Number 2 pp 119–128 (2002)DOI: 10.1002 /kpm.140

Copyright # 2002 John Wiley & Sons, Ltd.

enabling better real-time communication. An idealapproach would combine these with the ability tocreate knowledge content that can be either for-mal or informal in nature, in a rapid, real-timemanner. We will call this Rapid KnowledgeConstruction (RKC). In order for an RKC approachto be effective, it must be flexible enough torespond to particular situations and people, butin addition, both disciplined and capacious enoughto capture and link material on the fly. Artifactsalways shape the social and material context ofuse; in an RKC approach, the methodology andsupport tools must facilitate rapid reconfigurationfor different contexts (Figure 1). The context in turninforms the application (tools+methodology), in acontinuous cycle.

An RKC project can have many goals, time-frames, and audiences. For example, a projectmay involve a few people in a single workgroup,or many people distributed across a number ofgroups. Sometimes the target group and taskbenefit from a highly structured approach, becauseof external or contextual factors, whereas sometarget groups and tasks require a more explora-tory, less constrained approach. Some RKC pro-jects hold developing a body of content forsubsequent use as their primary goal, while otherprojects may have as much to do with developingbetter understanding, communication, or otherskills among the project group in the process ofdeveloping the content. An RKC approach shouldbe flexible enough to accommodate goals any-where along these continuums, and to move backand forth between emphases as needed. Theknowledge captured and stored in an RKCrepository should be available for reuse by newproject teams or efforts, in such a way thatindividual knowledge elements can be easilydiscovered, accessed, and recombined for newpurposes — in other words, recontextualized. Tosummarize, RKC as just defined is exploring oneof the most difficult parts of the KM technologydesign space — representations and user interfaceswhich are ‘lightweight’ enough to enable situated,

real-time capture, but which also sustain reuse —two goals which are often in tension with eachother (Buckingham Shum et al., 1997).

We now introduce several empirical and con-ceptual strands of work which inform ourapproach:

$ Cognitive overheads of capturing structuredinformation

$ Perspective making and taking$ Boundary objects between communities of

practice$ Memory as narrative (re)construction

We shall revisit these in the discussion, to showhow they are engaged by Compendium.

Cognitive overheads of capturing structuredinformation

All systems seeking to store information mustconfront the capture and reuse bottlenecks: who isgoing to enter the material and in what form toassist reuse? In our case, we need to capture andstructure ideas emerging from discussions inmeetings; this is extremely hard to do automati-cally, so human structuring of material is required.Structure is a double-edged sword: while it canusefully focus attention on important concepts,sharpen thinking, and yield computable represen-tations, there is ample evidence that many systemsfor storing and sharing human-structured infor-mation fail because the envisaged beneficiaries ofthe system simply do not have the motivation ortime to invest in codifying resources to reach acritical mass. We draw sobering lessons fromanalysing design rationale systems (BuckinghamShum and Hammond, 1994; Buckingham Shumet al., 1997), plus evidence relating to groupware(Grudin, 1994) and hypertext more generally(Conklin and Begeman, 1988). The lesson for KM(summarized recently by Shipman and Marshall,1999), is that systems requiring human-encoded,formalized, sharable information often expect toomuch of their users. A detailed account of howthe Compendium RKC approach in this paperaddresses this issue is presented in Selvin’s (1999)analysis, which we now extend with a number ofadditional concepts.

Perspective making and taking

Boland and Tenkasi (1995) highlight the impor-tance of what they term perspective making andperspective taking in ‘communities of knowing’.‘Perspective making is the process whereby a

Figure 1 Tools and methods are shaped by, and in turnshape, their context of use

CASE STUDY Knowledge and Process Management

120 A. M. Selvin and S. J. Buckingham Shum

community of knowing develops and strengthensits own knowledge and practices’ (p. 356). It is thisprocess that underpins the building of a com-munity’s identity: their basic assumptions, goals,terminology, stereotypes and modes of discourse.‘Perspective taking’ refers to the process of tryingto engage with another community’s perspective.This can be difficult when their respective ‘ways ofknowing’ assume radically different, strongly heldagendas. An implication of Boland and Tenkasi’sanalysis is the need for tools to assist in perspec-tive management.

Boundary objects between communities ofpractice

Wenger (1998) presents a detailed account of theconcept of the community of practice, of whichBoland and Tenkasi’s ‘communities of knowing’may be considered a specialization. Of particularinterest to us is the role that technologies couldplay in helping knowledge capture and reusewithin and across communities. Wenger reflectson two ways in which boundaries betweencommunities can be bridged. First, brokers areindividuals capable of speaking the languages ofmultiple communities, and coordinating and align-ing activities within each sufficient to achieve auseful degree of mutual understanding. Second,material or conceptual representations may serveas boundary objects (Star and Greisemer, 1989).‘Boundary objects are objects which are bothplastic enough to adapt to local needs and theconstraints of several parties employing them, yetrobust enough to maintain a common identityacoss sites’ (Star and Greisemer, p. 393). Examplesof material boundary objects that are used tocoordinate and mediate between diverse commu-nities include geographical maps, architecturalplans, and software specifications. To this, Wengerwould add language and practices as social/conceptual forms that both divide and bondgroups.

Memory as narrative (re)construction

There are two related threads that we wish tohighlight: conceptions of memory, and narrativemodes of sharing and sensemaking. First, Bannonand Kuutti (1996) argue that the term organizationalmemory is often used implicitly to mean a reposi-tory based on a ‘memory as bin’ metaphor,whereby material is unproblematically added andextracted. They point out that the cognitive andsocial sciences show that ‘memory as reconstruc-tion’ is a better model — material is not simply

retrieved from a ‘bin’, but is reconstructed in thecontext of our understanding of the world, whois asking, and for what purpose. The task fororganizational memory design is thus better con-ceived as the provision of resources for theconstruction, reconstruction and negotiation of infor-mation (e.g. over ‘what happened’, and ‘what thisnow means’).

Taking us to the second theme, Bannon andKuutti emphasize the important role that ‘talk’ and‘narrative’ seem to play in the construction ofmeaning. There are two senses in which narrativeseems to have a role to play in the construction ofmeaning. First, there is ‘the story as artifact’ — aconstructed account of past incidents in the lifeof the group, or of its members. Through stories,members share the latest news, show off expertise,seek advice and compare notes, in the processmaintaining their standing and claim to member-ship within that group (Orr, 1990). Second(although clearly not distinct from ‘stories asartifacts’), there is narrative as a mode of cognitionin contrast to logical/analytical cognition. Narra-tive in this sense is used to make sense of theworld by casting and recasting the past in rela-tion to present situations. Narrative theory isbeing drawn upon in KM-related research into‘sense-making’ (e.g. Dervin, 1983; Weick, 1995).The discourse-oriented representations used inCompendium are resources for (re)constructingthe ‘story’ of a project: what questions were asked,what answers considered, with what kinds ofrationale?

COMPENDIUM AS AN RKC APPROACH

The set of techniques which represent the Compen-dium approach are designed to be adaptive to aspecific context, and mobile along the tool/methodcontinuums of Figure 1. They revolve around agraphical hypermedia system1 for the develop-ment and application of (1) question-orientedtemplates which serve as semiformal ontologies tostructure the subject matter of a particular project(Figure 2), and (2) a set of metadata tags that can beassigned to any concept in the database (Figure 3).A hallmark of the approach is the ability to movebetween formal and prescribed representationsand informal, ad hoc communication, incorporatingboth in the same view if that is helpful to the

1In the examples presented in this paper, the tool is GDSS Inc. ’sQuestMap2 product <www.gdss.com>. Verizon is currentlydeveloping a Java-based product with more comprehensivesupport for the Compendium methodology.

Knowledge and Process Management CASE STUDY

Rapid Knowledge Construction 121

participants. Hypertext nodes and links can beadded either in accordance with templates or in anopportunistic fashion.

Compendium, first developed in 1993 as anapproach to aid cross-functional business processredesign (BPR) teams, has been applied in severaldozen projects in both industry and academicsettings (Selvin, 1996; 1998a; Selvin and Sierhuis,1999a,b). Its origins lie in the problems attendingteams working over weeks or months to designbusiness processes: keeping track of the plethora ofideas, issues, and conceptual interrelationshipswithout needing to sift through piles of easelsheets, surfacing and tracking design rationale, andstaying on track and ‘bought-in’ to the project’soverall structure and goals. The key feature of theearly approach was the combination of an Issue-Based Information System (IBIS) concept-mappingtool (Conklin and Burgess Yakemovic, 1991), whichsupported informal and exploratory conversationand facilitation, with a structured modelingapproach (Selvin, 1999). This allowed teams tomove along the spectrum of formal to informalcommunication as well as prescribed to sponta-neous approaches, as their needs dictated. It also

let them incrementally formalize data (Shipmanand McCall, 1994) over the life of the project.

As the approach was tested and refined over thecourse of several years, additional tools andmethods were added that facilitated RKC. Theearly use of the IBIS tool, while powerful, had anumber of limitations. First, often content alreadyexisted in the form of documents or other textualmaterials (email messages, web pages, etc.), and itwas not efficient to manually retype the material inorder to transform it to hypertext nodes and links.A tool was developed that analyzed documentcontent and metadata, such as paragraph styles,and automatically generated typed nodes andlinks in the hypertext database. Participatingteams wanted output in customized documentforms, not just the outline format that the hyper-text tool provided, so a tool was written togenerate multiple document formats suited tothe representational preferences of the audience(Selvin and Buckingham Shum, 1999). This wasexpanded to pictorial/graphic representations,such as data flow diagrams and even ad hocuser defined representations. Further extensionsmet teams’ desire for the rapid publication of

Figure 2 A compendium question template representing the concerns of a particular stakeholder group (application in Figure 4)



Compendium meeting minutes, analyses, textualversions of models, etc. on the corporate intranet.

VERIZON Y2KCP CASE STUDY

The following case study illustrates how Compen-dium has worked as an RKC approach. Thecontext was a highly time constrained, organiza-tionally pressurized effort to conduct an enter-prise-wide risk assessment for a Year 2000Contingency Plan (Y2KCP). The Y2KCP projectcomprised an intensive analysis effort to identifythose Verizon resources that would cause thehighest degree of risk if they were to fail. Theeffort was to look at the five ‘core businessprocesses’ of the telephone company as the outsideworld — customers, regulators, investors —perceived them. These processes were Call Com-pletion, Ordering and Provisioning, Maintenanceand Repair, 911 (emergency calls), and Billing.Cross-functional analysis teams were to be formedconsisting of members from many departments.The output was to be used both as the contentof a Corporate Contingency Plan document thatcould be given to external stakeholders, and as avalidation check against the departmental-level

contingency plans generated in an earlier phaseof the project.

Prior to the start of the work described in thispaper, a team of external consultants had spentseveral weeks canvassing various executives forparticipants and holding meetings attempting tokick off the analysis process. When these kickoffmeetings proved difficult and contentious as topurpose and process, the Compendium team wasasked to help facilitate and provide a structure anddatabase for the rest of the effort. Sessions wereheld in meeting rooms at Verizon locationsthroughout the northeast USA as well as over theVerizon network using a virtual team approach.This included the use of Microsoft NetMeeting22

and phone conferencing. About 60% of the meet-ings were face-to-face and 40% used the virtualteam approach. The project was conducted inspring 1999, with approximately 60 participants.

Initiating the Compendium process

Use of Compendium began with facilitated colla-borative sessions with project staff to clarify whatthe aims of the risk assessment were. Goals,

Figure 3 Optional metadata tags added to the content of a node, assisting subsequent harvesting and analysis of elements

2NetMeeting is a trademark of the Microsoft Corporation.



expectations, and specific analysis outcomes werediscussed and modeled in real time. Out of thosesessions came preliminary templates (e.g. Figure 2)for the process analysis and risk assessment com-ponents of the analysis effort. The templates werethen applied in an opening session with the CallCompletion team (Figure 4).

Feedback from the team and reflections afterthe session prompted refinement of the templates,which continued in subsequent sessions with theremaining groups. As the templates improved, theeffort picked up speed. After three or four ses-sions, the process moved into a consistent mode:introduce the project, problem, and approach;do a high-level process flow on paper or white-board; enshrine the agreed-on high-level flow inCompendium software tools; apply the processtemplate to identify which resources (people,computer systems, network equipment, buildings)pertained to each subtask; when all the resourceshad been identified attach the risk assessmenttemplate to each resource so that the team couldassess the operational, competitive, and legal risksattendant to that resource for that task.

Rapid construction of the contingency planninganalysis

As the teams got deeper into the analysis, theywere able to leverage each other’s work. Forexample, several of the teams identified the samesystems, roles, and tasks, often adding moredescription or definition in the process. Compen-dium supported teams in systematically usingonly one node to represent a resource (people,systems, equipment, locations, etc.), for example,‘the WFA system’ (as illustrated in Figure 5). Thus,teams could leverage the work that other teamshad done by reusing the same node in multi-ple contexts. ‘Transclusive hyperlinking’ (Nelson,1987) meant that each team could add correc-tions or descriptive information to a resource, andknow that it was immediately updated in everyprocess map and risk assessment where itappeared. In addition, since the software tooldisplayed ‘libraries’ of the resources, the teamscould save time by simply copying icons into theircurrent analysis map. The result was that onecould find all the maps in which a resourceappeared, helping to answer key questions such

Figure 4 Early session with Call Completion team (using the template in Figure 2)



as ‘in what contexts is this factor relevant?’ and‘who are the stakeholders?’

Participants identified which resources wereused in each work step of the business processes.Following this, the resources were rated in termsof the competitive, legal, and operational riskshould that resource fail. Resource/task combina-tions rated as ‘high risk’ were then subjected to aseparate contingency planning effort. Due to thetight timeframes and high priority for the project,there was no time available for the more explora-tory discussion of the many issues involved thatCompendium has supported in other contexts(Selvin and Sierhuis, 1999a); the facilitator focusedthe discourse on completing the analysis tasksassociated with the visual maps displayed inmeetings.

By using a visual map continuously created inreal time by the facilitators, with all the contentprovided by team members, validation of theknowledge database’s content could occur as itwas being created. Participants saw the maps ofhypertext nodes and links being created, andinformed the facilitator if there were inaccuraciesor omissions. A second ‘round’ of validation

occurred when documents and diagrams werecirculated. The project’s sponsors required thatthe output be customized for different audiencesin various document and diagrammatic forms.Compendium software was used to generatetwenty-five different textual documents and morethan 500 process diagrams for immediate feedbackand review.

Knowledge reuse

A critical indicator of the efficacy of an RKCapproach is the ease with which knowledgeelements can be retrieved and recombined in newways, for new purposes. The case study discussedin this paper has featured two subsequent reuseinstances in the months following the originalY2KCP work. Following the completion of theContingency Planning effort, a new team, com-posed of people who had not been involved in theY2KCP effort, began overlaying the Y2KCP con-tent with an analysis of which resources and taskshad been covered by Y2K-related Business ProcessTesting. The Business Process Testing team madeuse of a variety of reuse mechanisms (Selvin,

Figure 5 Use of the WFA node in two different teams’ maps. One can display a menu of all contexts in which a node has beenused; selecting a context highlights the relevant node in its context of use



1998b) to accomplish this work. Later, an eBusi-ness Strategy team used the process analysis tooverlay internal eBusiness initiatives and asso-ciated templates onto the core business processes(Figure 6).

DISCUSSION

Effective, on-the-fly construction of knowledgeresources does not come ‘for free’ — lowering theeffort invested at the capture stage (e.g. simplyvideo recording all meetings, or taking conven-tional minutes), merely leaves more work sub-sequently to support reuse. Naturally, we wantquality knowledge resources for minimal effort,but while information extraction and other tech-nologies for analysing unstructured sources con-tinue to develop, they do not support the groupsensemaking processes that Compendium assiststhrough its shared, visual maps. Moreover, therewere pragmatic factors to consider in this casestudy: what tools were available now to tackle Y2Kanalysis and planning across diverse stakeholdergroups? Compendium’s approach is to tackle the

cost – benefit tradeoff by investing effort in qualitycapture by a facilitator. Individuals exposed toCompendium then learn to use it themselves, ormay receive further training on short courses.

The approach draws on several conceptualframeworks introduced earlier, illustrating theirpotential for framing KM technologies. Compen-dium assists the making and taking of perspectives bymaking perspectives tangible (through reusabletemplates that set an agenda), which then enable‘taking’ dialogues to occur (stakeholder commu-nities encounter, and can understand each others’perspectives within a common environment andlanguage). Compendium makes heavy use ofvisual boundary objects to help stakeholder groupsand analyst/facilitators (the brokers) see howother groups view a problem, and emphasizes theparticipatory evolution of modelling schema andcontent from the language of a given community ofpractice. In Star and Greisemer’s terms, Compen-dium’s representations are robust enough to mean-ingfully capture discussions for subsequent reuse,yet plastic enough to allow negotiation over themeaning of elements of common interest (usinggraphical argumentation structures if desired, or

Figure 6 An eBusiness process testing analysis (BPT clusters?) overlaid onto the original Y2KCP process model (B2.3.3 CarrierIn-Collect Process)



just visually mediated talk). The maps representmultiple perspectives at a granularity and level ofabstraction that is acceptable to all stakeholders(preserving local control over detail), and can alsotransform one community’s perspective into moreappropriate formats (boundary objects) for others,thus assisting their sense-making (see Selvin andBuckingham Shum, 1999, for details of ‘represen-tational morphing’). Incremental formalization ofnodes is possible on a ‘just in time/just as needed’basis in order to tackle the capture challenge ofcognitive overhead. Finally, the ability to movesmoothly between highly disciplined modelling,semi-structured argumentation, and informalnotes/discourse makes provision for the co-existence of conflicting views or uncertainty,consistent with a view of organizational memoryas negotiated, constructed and situated with respect tothe codification and interpretational context.

To summarize, this paper has described Com-pendium, which addresses a number of key issuesfor the successful introduction of RKC KM tech-nologies into work practice.

(1) Customization for different use contexts —Y2KCP templates and metadata;

(2) Integration of formal and informal communi-cation, and

(3) Integration of prescribed and ad hocrepresentations — due to the well understood,highly constrained nature of Y2KCP, there wasless informal communication and representa-tion than in other case studies (Selvin andSierhuis, 1999a);

(4) Validation and cross-referencing of the reposi-tory ‘on the fly’ at the point of entry (by theY2KCP facilitator and the teams’ monitoring ofhis work);

(5) Conversion of organizational documents/emails into a hypertext database — see Selvinand Buckingham Shum (1999) and Selvin et al.(2001) for examples from this and otherdeployments of the approach;

(6) Conversion of hypertext databases into organi-zational document formats — textual docu-ments and process diagrams generated fromY2KCP maps

The Y2KCP project and subsequent reuses ofthe knowledge gathered for new purposes illus-trate how the Compendium approach can serveas an effective Rapid Knowledge Constructionapproach. The teams described were able to getup to speed quickly and conduct a complexanalysis in a tight timeframe, gather and validatea large body of knowledge, produce documents

and artifacts in a variety of forms for differentaudiences, build on each other’s work, and adaptthe content for new purposes. Compendium’sreusable templates and hypermedia map elementsdemonstrate the power of visualizations for estab-lishing a common language. In addition, granular,hypertextual linking of important concepts in aproject enables the recontextualization of concepts,while maintaining links back to source documents.

The Compendium approach is currently beingused and explored for similar purposes by avariety of organizations, including the NASAAmes Research Center, Wisconsin Public Televi-sion, The Open University’s Knowledge MediaInstitute, and the Center for Creative Leadership.This community, now networked via the virtualCompendium Institute,3 is already demonstratingthe fruit of collaboration with recent technical andconceptual extensions (Selvin et al., 2001) to thework reported here. We welcome approaches fromothers who wish to work with us.

REFERENCES

Bannon LJ, Kuutti K. 1996. Shifting perspectives onorganizational memory: from storage to active remem-bering. Proc. HICSS’96: 29th Hawaii International Con-ference on System Sciences, Vol. III – CollaborationSystems and Technology, Hawaii, January. IEEE Com-puter Society Press: Washington, DC; 156–167.

Boland RJJ, Tenkasi RV. 1995. Perspective making andperspective taking in communities of knowing. Organ-ization Science 6(4): 350–372.

Buckingham Shum S, Hammond N. 1994. Argumentation-based design rationale: what use at what cost?International Journal of Human–Computer Studies 40(4):603–652.

Buckingham Shum S, MacLean A, Bellotti V, HammondN. 1997. Graphical argumentation and design cogni-tion. Human – Computer Interaction 12(3): 267–300.<http://kmi.open.ac.uk/tr/papers/kmi-tr-25.pdf>

Conklin J, Begeman ML. 1988. gIBIS: A hypertext tool forexploratory policy discussion. ACM TransactionsonOffice Information Systems 4(6): 303–331.

Conklin J, Burgess Yakemovic KC. 1991. A process-oriented approach to design rationale. Human –Computer Interaction 6(3&4): 357–391.

Dervin B. 1983. An overview of sense-making research:concepts, methods and results. Annual Meeting ofthe International Communication Association, Dallas,TX, May. <http://communication.sbs.ohio-state.edu/sense-making/art/artdervin83.html>

Grudin J. 1994. Groupware and social dynamics: eightchallenges for developers. Communications of the ACM37(1): 92–105.

Nelson T. 1987. Literary Machines (Ed. 93.1). <http://www.eastgate.com/catalog/LiteraryMachines.html>

Orr J. 1990. Sharing knowledge, celebrating identity:war stories and community memory in a service

3Compendium Institute: www.CompendiumInstitute.org



culture. In Collective Remembering: Memory in Society,Middleton DS, Edwards D (eds). Sage Publications:Beverly Hills, CA.

Selvin A. 1996. Leveraging existing hypertext function-ality to create a customized environment for teamanalysis. Proceedings of the 2nd International Workshopon Incorporating Hypertext Functionality Into SoftwareSystems.

Selvin A. 1998a. Facilitating electronically: using tech-nology to help Maria. The Facilitator (Special Issue onAutomated Meeting Support) September.

Selvin A. 1998b. Supporting granular reuse of knowl-edge elements in an organizational memory system.Proc. 7th International Workshop on Hypertext Function-ality (Organizational Memory Systems & HTF), Helsinki,12–13 December. University of Oulu, Dept Com-puter Science Technical Report Series A 29 (November99).

Selvin A. 1999. Supporting collaborative analysis anddesign with hypertext functionality. Journal ofDigital Information 1(4). <http://jodi.ecs.soton.ac.uk/Articles/v01/i04/Selvin/>

Selvin A, Buckingham Shum S, Sierhuis M, ConklinJ, Zimmermann B, Palus C, Drath W, Horth D,Domingue J, Motta E, Li G. 2001. Compendium:making meetings into knowledge events. KnowledgeTechnologies 2001, March, Austin TX. <http://www.CompendiumInstitute.org/compendium/papers/Selvin-KT2001.pdf>

Selvin A, Sierhuis M. 1999a. Argumentation in differentCSCA project types. Workshop on Computer-SupportedCollaborative Argumentation, Conference on Computer-Supported Collaborative Learning, Stanford, CA, 12–15

December 1999. <http://kmi.open.ac.uk/sbs/csca/csc199>

Selvin A, Sierhuis M. 1999b. Case studies of ProjectCompendium in different organizations. Workshop onComputer-Supported Collaborative Argumentation, Con-ference on Computer-Supported Collaborative Learning,Stanford, CA, 12–15 December. <http://kmi.open.ac.uk/sbs/csca/csc199>

Selvin AM, Buckingham Shum SJ. 1999. Repurposingrequirements: improving collaborative sensemakingover the lifecycle. Profess’99: International Conferenceon Product Focused Software Process Improvement,Oulu, Finland, 22–24 June 1999, 539–559. <http://www.inf.vtt.fi/pdf/symposiums/1999/S195.pdf>

Shipman FM, Marshall CC. 1999. Formality consideredharmful: experiences, emerging themes, and direc-tions on the use of formal representations in inter-active systems. Computer Supported CooperativeWork 8(4): 333–352. <http://bush.cs.tamu.edu : 80/yshipman/cscw.pdf>

Shipman FM, McCall R. 1994. Supporting knowledge-base evolution with incremental formalization. Proc.ACM CHI’94: Human Factors in Computing Systems,Boston, MA. ACM Press: New York; 285–291.

Star SL, Greisemer J. 1989. Institutional ecology, ‘transla-tions,’ and coherence: amateurs and professional inBerkeley’s Museum of Vertebrate Zoology, 1907–1939.Social Studies of Science 19: 387–420.

Weick KE. 1995. Sensemaking in Organizations. SagePublications: Thousand Oaks, CA.

Wenger E. 1998. Communities of Practice: Learning,Meaning, and Identity. Cambridge University Press:Cambridge.



rapid knowledge construction: a case study in corporate contingency planning using collaborative...

Documents