International Conference on Computer Systems and Technologies - CompSysTech’09
Keeping Artifacts Alive: Towards a Knowledge Management System Hilda Tellioğlu Abstract: The paper investigates first the role of artifacts in a cooperative work environment. It shows the different types of artifacts used in work groups. Then, it stresses the characteristics of common artifacts and shows their lifecycle. One way to support knowledge management in organizations is to make the useful information captured in artifacts and additionally annotated or linked information permanent. The paper finally introduces two possible mechanisms to achieve this: ordered exchange vs. shared space approaches. Before concluding, it describes briefly a knowledge management system with snapshots. Key words: Artifacts, CSCW, knowledge management.
INTRODUCTION Artifacts are the main resources hosting information, triggering action, documenting processes, and maintaining knowhow and experiences. For this reason artifacts are central for production or to provide services by any kind of organization. They are created, used, modified, communicated, sometimes stored and sometimes destroyed by actors or systems established in organizations. Storing people’s knowhow, experiences gained in projects, enabling retrieving of relevant data when needed are used in organizations for several purposes. Some researchers have studied the role of artifacts in cooperative work. Frameworks to address specific aspects of coordinative practices are developed, e.g., organizational memory, common information spaces, workflow systems, coordination mechanisms, mechanisms of interaction, boundary objects [10]. Others developed further concepts like situated and predefined coordination [6] for coordinated work environments [12]. Definitions of these concepts are analyzed and validated in ethnographic case studies. New systems have been designed to support cooperative work based on the analysis of artifacts. However, there is no study on artifacts as central entities for knowledge management in organizations and no one investigated artifacts as such. This paper tries to analyze the lifecycle of artifacts in a cooperative work environment. It describes on an example how to follow central common artifacts in a networked setting and that artifacts have a lifetime before they disappear from the use field. Knowledge management in an organization is mainly based on the information saved in artifacts. If artifacts disappear, the hosted knowhow and experiences disappear as well. The paper shows how to keep artifacts and with them the information that is included or attached to them alive. It uses the concept of snapshots for knowledge management [11] and tries to discuss mechanisms of maintaining artifacts as essential entities for knowledge management. In the next section different types of artifacts are described to show the wide range of their use in the context of collaborative work setting. After stressing the characteristics of common artifacts, their lifecycle are shown. To support knowledge management in organizations, the paper introduces the possible mechanisms: ordered exchange vs. shared space. Finally it describes briefly a knowledge management system with snapshots before concluding. ARTIFACTS IN COOPERATED WORK An artifact is a permanent symbolic construct and plays a very important role in the coordination of cooperative work. The symbolic artifacts act as mediators of the coordination. They are used to clarify ambiguities and to settle disputes. Artifacts mediate articulation work by acting as an intermediary with a specific material format between actors. “Instruments, signs, procedures, machines, methods, laws, forms of work organization” are examples of artifacts [4, p.26f]. Furthermore “an object can be a material thing, but it can also be less tangible (such as a plan) or totally intangible (such as a
International Conference on Computer Systems and Technologies - CompSysTech’09
common idea) as long as it can be shared for manipulation and transformation by the participants of the activity”. In a usual work environment actors interact by using several artifacts. Where an artifact is placed, when it is accessed, by whom it is modified, what the exact modifications are, etc. do have implications to work processes. Lundberg and Sandahl try to investigate artifacts’ properties in work by analyzing and identifying how artifacts’ peripheral properties become common resources within a community of practice and how artifacts are active within a work environment [5]. Artifacts may coordinate activities by their visibility or by their arrangements. Wartofsky defines three types of mediating artifacts [14], which has been further developed by Engestrom [2] into a three-level hierarchy. “Primary artifacts are tools used directly in production to mediate the relationship between the subject and object of activity; secondary artifacts are representations of modes of action – models – used to preserve and transmit skills in the production and use of primary artifacts; tertiary artifacts are imaginative or visionary and give “identity and overarching perspective to collective activity systems”” [3, p.3]. This scheme has been further developed within activity theory [2] [1] and includes the following artifacts: What-artifacts contribute a means of achieving the object, how-artifacts contribute to understanding how to achieve the object, whyartifacts motivate achievement of the object, where-to-artifacts motivate evolution of all elements in the activity system. Specialized artifacts [7, p.162] like time tables, schedules, catalogues, classification schemes for large repositories, etc. help reducing the complexity of articulation work and alleviating the need for ad hoc deliberation and negotiation. Material artifacts are accessible to all involved in a shared work process [8]. The location of a material artifact includes some relevant information. Some of the actors can probably make sense of it and some others not. One can also see the history on a material artifact, the history of past work as well as the contributions of different actors. Representational artifacts are the immediate objects of the work [9]. They are objectifications of things-to-come and of things-in-the-process-of-becoming. Representations are not real, are local and temporary, and are “conventionalized practices based on rules of mapping and translation between representation and the object that is represented” [9, p.15]. Coordinative artifacts are material artifacts that have a coordinative role in carrying out work practices. In this sense they are communication objects and persuasive [13]. They help actors in several ways: They create a common understanding of an idea or task. They enable talking about an idea in a rich way. They remind principles, approaches and methods applied, questions that are still open. They also help keeping track of activities and materials [8]. Some artifacts that are mainly used for coordination purposes contain work plans. They include work to do, project phases, how to proceed in a specific project phase, material to collect or create, methods defining rules and conventions within the work group, illustrations like sketches, images or photos to explain the idea, references to material to look for, names of actors responsible for certain tasks. As boundary objects coordinative artifacts are accessed and modified by all responsible actors. They enable crossing organizational and professional boundaries many times. All decisions made in a project can be captured by using common artifacts. As several actors have access to these shared artifacts the concurrency needs to be monitored and a version control mechanism must be provided, electronically or conventionally. It is very common that actors add annotations to artifacts they use by circling certain areas, adding notes, marking a certain part with a marker or with a post-it including additional instructions or important notes to be communicated among the actors. This type of annotations makes artifacts multilayered. Multilayered artifacts “facilitate coordination between activities (and the people who are responsible for them). They, for instance, provide a collective or
International Conference on Computer Systems and Technologies - CompSysTech’09
individual space for experimentation and change” [8, p.10]. That implies that artifacts are interrelated to work activities. There are several types of coordinative artifacts: Traces are artifacts that indicate aspects of past activities of coordinating actors. Templates are artifacts that specify the properties of the result of individual contributions, like product standards, drawings, style sheets etc. Maps are artifacts that specify interdependencies of tasks or objects in a cooperative work setting, like organizational charts, classification schemes, taxonomies etc. Scripts are artifacts that specify a protocol of interaction in view of task interdependencies in a cooperative work setting, like checklists, productions schedules, office procedures, bug report forms etc. Characteristics of common artifacts can be described as follows: A common artifact implements the basic functionality to fulfill the requirements. Its structure and operation is predictable. Actors need to see any time what others are doing and a common artifact offers this required overview of the cooperative work. Implicit communication within a team is enabled by common artifacts. A common artifact serves as a template representing a limited model of the work to be done by leaving enough space to be filled by its users. If templates are filled then they become records of the work. When several actors access the same artifacts these common artifacts must be identified and validated. “When accessing an artifact produced and submitted by somebody else, the actor who may need to retrieve it must be able to establish its particular identity. The item thus has to be named or otherwise identified so that a potential user will know ‘which’ it is” [9, p.19]. In this context conventions to name the items must be defined and communicated. “When accessing an artifact produced by somebody else, the actor retrieving it will also need to somehow assess its relevance, validity, veracity, etc.” (p. 19). Given these different types, we could identify several artifacts in our case study what we want to describe in the next section. THE LIFECYCLE OF ARTIFACTS We carried out ethnographic studies in a European STREP project called MAPPER (Model-based Adaptive Product and Process Engineering) (IST-016527)1 at three different industrial sites. One of them is KeyA that we refer in this paper to. KeyA is a company that produces car parts like gearshifts, head strains, and seat heating for automotive industry. It has several branches all over the world. The projects are multinational, involving different people from different branches depending, on the one hand, on people’s skills and knowledge, and on the other hand, on the production facilities at the site. We could observe several meeting and carry out in-depth interviews with some of the key actors involved in the projects investigated. The most important artifact for project management in our study was a to-do list used during regular meetings. A to-do list was a list of open issues related to an ongoing project. Each open issue was named, described, assigned to a person, got a deadline and a status (Figure 1). These lists had a long life, almost certain phases of a project, because they were used to report things done or issues that could not be solved in further meetings with top management or suppliers or even directly with customers. We could not find any to-do list of past projects that have already been finished or cancelled. Project members had their own to-do lists, usually handwritten, unstructured, messy and very specific. These individual lists were mainly used until the next meeting to remind the project members to things what they have to do. Some issues survived couple of meetings; some were removed after couple of days.
1
http://mapper.eu.org/
International Conference on Computer Systems and Technologies - CompSysTech’09
Figure 1. Computer-based to-do lists of project managers with general issues, responsible persons and deadlines (top) vs. handwritten to-do lists of meeting participants for individual use (bottom).
Because issue items disappeared when they were solved and were removed from todo lists, no one could keep track of solved issues, no one knew after a while what has happened to that issue, if there were any decision made, why they have been made and why certain mistakes happened related to that issue. The example above shows that artifacts have their lifetime. They disappear when they are not needed any more. And all the knowledge connected to them disappears as well. This is exactly the problem that organizations face nowadays and do not know how to deal with. MECHANISMS FOR MAKING ARTIFACTS PERMANENT To avoid the loss of very useful information hosted by artifacts like to-do lists, mechanisms of knowledge management must be established. The architecture of these mechanisms is based on the interaction modes applied by work groups in dealing with artifacts. As described before, artifacts of different kinds do play a very important coordinative and mediating role. The ways of exchange of (common) artifacts shape the mechanism that is needed to support, on the one hand, the interaction between cooperating actors or groups, and on the other hand, the knowledge management between all. In this sense, we need to distinguish between two different approaches: ordered exchange versus shared space. Table 1: Comparison of system properties of ordered exchange and shared space approaches. System properties Ordered exchange approach Shared space approach Description Artifacts move from one actor to Shared artifacts are located in another. common information spaces. Access onto common artifact Sequential Simultaneous and sequential Interaction paths 1:1, 1:n, n:1 n:m Temporal arrangements One actor occupies the artifact for Artifact is available for all actors all the a period of time. time. Restrictions by write access through lock mechanisms, not concurrent access, use of configuration management. Use of views Artifact is the view, otherwise a Artifact is not necessarily the view. new artifact must be created (for Views are necessary to filter data for the access by a specific actor, access by specific users. e.g., only on a part of the artifact). Access order Ordered access Unordered access Access initiator Actor or the system used deliver Actors are responsible to get and put artifacts to actors. the common artifact.
International Conference on Computer Systems and Technologies - CompSysTech’09
Ordered Exchange Approach The main property of this approach is that artifacts move from one actor to another. The access to the common artifact can only occur sequentially. Communication enables information exchange between distributed actors. For different purposes there are different types of communication patterns (one-to-one, one-to-many or many-to-one). These types occur sequentially within work practices initiated by actors or the system used. Artifacts exchanged are the view of the work-in-attention. They represent the work. They help actors to organize their work. At the same time, they are mechanisms of articulation work. They can also be used to create and maintain boundaries between communities of practice. Finally, views are interpreted by actors creating or using them. This depends also on actors’ disciplines, professions, knowledge and skill. Shared Space Approach In this approach artifacts are mainly shared. These common artifacts are located centrally and are available for read and write access of project members. By applying the appropriate locking mechanisms, concurrency controls or version controlling system, not only sequential but also simultaneous access to common artifacts are provided. Interactions between people using the common information space and accessing the common artifacts are characterized as many to many. In general, the access is unordered. It is the responsibility of actors to get or put the common artifact to the central repository. Knowledge Management with Snapshots If we can differentiate between these two approaches in a real collaborative work setting and capture the central artifacts needed in concrete work processes, the next step is to think about mechanisms of making artifacts permanent. This means to store not only the content contained in artifacts, but also the meta-level information about their authors, the date of creation and modification, the location, the size, the deadline and responsible person if it is a to-do list item, from and to fields if it is an email, etc. A snapshot-based knowledge management mechanism [11] creates a network of projects, people and artifacts including content- and meta-level information. Additional to this meta-level information, links between artifacts created by users, tags and comments given to artifacts to make more sense of them and ease their useful retrieve can be provided. If this is done in an implicit way without bothering users in their daily work, the value of this type of annotations and indexing are essential for further use of all information saved in artifacts. CONCLUSIONS In this paper, we presented our research on artifacts as key players in the coordination of cooperative work. There are mediating artifacts at different levels and complexity (primary, secondary, tertiary); what-, how-, why- and where-to-artifacts; specialized, material, representational and coordinative artifacts, which can be multilayered. Traces, templates, maps, and scripts fulfill different functions and used differently. In a cooperative setting, the focus is on common artifacts, like to-do lists, as we could study in our ethnographic case study. A common artifact can be used to exchange work in progress implicitly, to support articulation work, e.g., by representing work carried out, to point out possible and actual gaps in coordinating dependencies between tasks, to communicate the to-dos explicitly, to assign tasks to persons, to define and refine work to do, etc. Common artifacts host valuable information about processes, people’s knowledge, experiences, and habits, conventions and decisions made. If they are actively used, this information is also available for people using them. Unfortunately, artifacts have their life times. If they disappear the information embedded in them disappears too. This causes a gap in accessing knowledge and experiences of an organization.
International Conference on Computer Systems and Technologies - CompSysTech’09
Based on the use of artifacts by cooperating people, the ordered exchange and shared space approaches help us to design the right work environment for work groups. This distinction makes the use of temporal arrangements and views more transparent and accessible. The concept of snapshots for knowledge management [11], on the one hand, demonstrates the usefulness of these two approaches for making artifacts permanent, and presumes their existence on the other. However, there are some consequences of the artifact-centered approach. Using common artifacts help reduce coordination effort in a team, e.g., by being part of the product to be developed or integrated in the product and therefore by making additional communication or exchange obsolete. Can everything be captured in artifacts? Can the exchange of artifacts replace the face-to-face interaction of people working together? Designers of systems to support cooperative work need to be aware of the importance of personal contact and informal exchange, trust and confidence issues in a group, body language and so forth in communication and consider these important factors in the design of their systems. REFERENCES [1] Collins, P., Shukla, S., D. Redmiles. Activity theory and system design: A view from the trenches. Computer Supported Cooperative Work: The Journal of Collaborative Computing 11 (1-2), 2002: 55-80. [2] Engeström, Y. Perspectives on Activity Theory. Cambridge University Press, Cambridge, Chapter in Innovative learning in work teams: Analyzing cycles of knowledge creation in practice, 1999: 377-404. [3] Guy, E. S. Patters as artifacts in user-developer collaborative design. In Proceedings of ECSCW 2003. [4] Kuutti, K. Activity Theory and Human-Computer Interaction. The MIT Press, Cambridge, Massachusetts, USA, Chapter in Activity Theory as Potential Framework for Human-Computer Interaction Research, 1996: 17-44. [5] Lundberg, N., T. I. Sandahl. Artifacts in work practice. In IRIS 1998. [6] Lundberg, N., H. Tellioğlu. Understanding Complex Coordination Processes in Health Care, Scandinavian Journal of Information Systems, Vol. 11(2), University of Aalborg, Denmark, 1999: 157-181. [7] Schmidt, K., C. Simone. Coordination mechanisms: Towards a Conceptual Foundation of CSCW Systems Design. Computer Supported Cooperative Work 5, 1996: 155-200. [8] Schmidt, K., I. Wagner. Coordinative artifacts in architectural practice. In Cooperative Systems Design: A Challange of the Mobility Age, Blay-Fornarino, Ed., 2002: 257-274. [9] Schmidt, K., I. Wagner. Ordering systems. Coordinative practices and artifacts in architectural design and planning. OrdSys Journal of Computer Supported Cooperative Work, 2004. [10] Tellioğlu, H. Cooperative Work, Habilitationsschrift, Technische Universität Wien, Fakultät für Informatik, Wien, 2007. th [11] Tellioğlu, H. Knowledge Management with Snapshots, Proceedings of the 10 IFIP Working Conference on Virtual Enterprises, PROVE’09, 2009. [12] Tellioğlu, H. Modeling Coordinated Work: Definition and Application of the Model “Coordinated Work Environment”, The Journal of Supercomputing. An International Journal of High-Performance Computer Design, Analysis, and Use, Vol. 24(2), 2003: 161-171. [13] Wagner, I. Persuasive artefacts in architectural design and planning. In Proceedings of CoDesigning 2000, Nottingham, 11-13 September 2000: 379-390. [14] Wartofsky, M. W. Models: Representation and the Scientific Understanding. D. Reidel Publishing Company, Boston, USA, 1979.
ABOUT THE AUTHOR Assoc. Prof. Hilda Tellioğlu, PhD, Institute of Design and Assessment of Technology, Multidisciplinary Design Group, Vienna University of Technology, Phone: +43 1 58801 18716, Е-mail:
[email protected].
