Figure 1: Growth of Communities within Enterprises
The goal of many organizations is to use their information resources
effectively to identify new opportunities and create new products and services.
This often starts with enterprises looking at ways to gain benefit from
the information capital in their Intranet sites. In this respect evolution
to knowledge sharing in many enterprises is expected to follow the path
shown in Figure 1. Enterprise workgroups can be formed to combine the explicitely
stored knowledge in Intranets and their tacit knowledge to develop new
knowledge. These communities can evolve from relatively informal exchanges
to focused knowledge sharing processes with identified goals, agreed upon
governance structures and supported by knowledge sharing tools.
Figure 1: Growth of Communities within Enterprises
Describing evolution like that shown in Figure 1 is difficult in terms of current paradigms. What is needed is a way to express community evolution in the context of knowledge development activities in organizational settings. This is made even more difficult as there is no generally accepted paradigm for knowledge sharing (McAdam, McCreedy, 1999) although knowledge sharing and management is now becoming almost universally accepted as a goal in most advanced computer installations. To many, knowledge management is a way of collecting information in easily accessible formats using Internet technologies and making them available through document management software. This virtually corresponds to the first level in Figure 1. Earlier work suggested a way of defining knowledge sharing actions based on Nonaka's model and the last level of Figure 1. Workspaces that support this level were found lacking as they assumed that fixed structures can be provided to users without the necessary process of forming communities. We have now extended earlier work to include community formation within our metamodel. The emphasis of the metamodel is on goal oriented communication in organizational settings and includes the ability to evolve from open communities to communities with agreed governance structures (Jones, 1997).
Our metamodel continues to be based on the abstract notion of shared workspaces that can dynamically evolve by spawning off new workspaces. Flexibility is achieved through the ability to create a new workspace, invite people into it, initiate new actions, and provide people with the support needed to carry out their work in a collaborative manner. Additional flexibility is now provided to allow independent workgroups to form by integrating community activities, information and knowledge sharing into what are here called Knowledge Intensive Work Activities (KIWA).
Step 1 - Create a workgroup or community,
Step 2 - Identify the group tasks and set up workspaces for each task, and
Step 3 - Organize the work of each workspace.
This tends to cover most communities. In practice many open communities do not go beyond step one. Some others may never evolve into complex work structures and thus miss step 2 and carry out all their work in the one workspace. Still others within enterprises follow the three steps. Evolution itself can follow different paths depending on initial conditions.
Community aspects
Central to knowledge intensive processes is the growth of effective communities. Effective communities bring in people with the needed expert knowledge and disseminating it through the right organizational channels. At the same time they must create an environment that supports community growth and participation (Furst, Blackburn, Rosen, 1999). An important aspect is that work in communities work practices should evolve. The usual evolution here is through orientation, trust building, commitment (Warketin and Beranek, 1999) and then implementation. To do this requires activities to include ways that allow their participants to vary their work practices.
We look at community evolution in a totally open way. It should be possible
to initiate groups, combine them, or split them up into subgroups. As an
example we look at the growth of a publications process. Figure 2 illustrates
a possible scenario for community evolution. It might start with a publications
group assembled by a publications manager, Benny. This group may have a
workspace that concentrates on the discussion of ideas, often proposed
by sponsors. Once an idea is found useful a project group may be formed.
In this case an editor is appointed and people appointed to the group.
Thus as step 1 Amy is appointed editor in project 1 and Thomas, Cleo and
Mary assigned to this group. Amy and Cleo still stay within the publications
group. A second project is then formed and so on. The other aspect of evolution
is the changing nature of the workgroups as they proceed from early orientation,
through goal formation, commitment and implementation (Warnekin and Beranek,
1999).
Community growth itself, however, must mirror community norms. There may be open and closed in the same enterprise and they must interact. If they are closed how does membership evolve. Furthermore within the same environment there may be a range of communities. In a teaching environment the group that sets assignments and distributes them is closed and organizationally controlled. At the same there may be groups of students who can set up their own workgroups to share their knowledge. Our metamodel thus includes constructs that support such community evolution within the one application structure.
Work within communities
Workgroups themselves may evolve to structures that involve a number
of workspaces like that shown in Figure 3. Knowledge intensive work activities
seen here as a network of connected work communities or groups, developing
and exchanging artifacts. Each activity is defined in terms of abstract
metamodel concepts defined earlier (Hawryszkiewycz, 1997). The workgroup
is organized through the definition of roles with specified responsibilities.
People are assigned to carry out the roles. Our goal is to develop a model
for growth that can start with an open community and evolve to a structure
like that in Figure 3.
Figure 3: Concepts to describe knowledge intensive work activities
The design of such systems requires a combination of good community practices within each activity, integrated into inter-activity processes through coordination algorithms that emphasize awareness and joint milestone formation rather than strict workflows. It also requires workgroup membership to be consistent with organizational culture.
Growth of knowledge intensive processes can follow many paths. Even within enterprises they can start as a totally open community with an idea that it later pursues in some organized way. Alternatively existing workgroups can initiate new tasks and create structured workgroups to follow them up. This then requires coordination mechanism to coordinate these groups. The coordination methods may themselves evolve.
Knowledge Sharing Processes
The important characteristic of knowledge creation is that it cannot be predefined but emerges. Generally it is found that such processes are opportunistic in nature (Dourish, 1999), result in islands of disconnected work activities, which nevertheless must be coordinated towards some common goal. Furthermore, activities evolve dynamically and sometimes rapidly as new situations arise. Evolution can include changes in participants, goals, and methods of collaboration as people begin to understand the collaborative tools and learn how to collaborate electronically. The evolution must be user driven and any collaborative tools should allow the users themselves to initiate such change and modify their processes. Thus it should be possible at any time for an activity participant to initiate some new actions and define events and milestones which may need follow-up actions.
The other aspects are the presentation of information in ways conducive to knowledge sharing combined with actions that facilitate interpretation and evaluation (Nonaka, 1994). This paper proposes that one way to support the activities found in knowledge intensive work is to use shared workspaces that can display all the actions in a single workspace and can dynamically evolve. Such integration of actions will allow people to see and interpret their work in the entire context. Flexibility is achieved through the ability to create a new workspace, invite people into it, initiate new actions, and provide people with the support needed to carry out their work in a collaborative manner. It allows new events to be defined dynamically as work evolves.
The metamodel centers on generic concepts for collaborative processes (Hawryszkiewycz, 1998) that include roles, activities, artifacts and actions. The metamodel centers on shared workspaces that have any number of roles that are occupied by people. Recently we expanded the metamodel to include the idea of workgroups and thus support the ability of groups to evolve independently and later combine into more complex coordinated structures. It also support scalability in the sense that independent workgroups can exist in the same system but gradually merge or intersect if needed.
The semantics are that a workspace represents an organizational recognized
activity, whose goal is to produce an organizationally recognized artifact.
Any number of people may be involved in the activity. The people are assigned
roles, which can take a variety of actions, both individually or in collaboration
with others. Usually the actions are combined into tasks that may be needed
to produce intermediate outcomes prior to the production of the activity
artifact. Each workspace can include a number of events that can lead to
message being sent to other workspaces, thus supporting emerging workflows.
Workgroups differ from workspaces in two major ways ñ separation of communities and management of growth. The general effect is that people in a workgroup can only participate in workspaces in that workgroup but not in workspaces of other workgroups. This is needed in situations where there is geographic separation with responsibilities vested in geographic teams, or in situations where one might be managing different client groups that require maintenance of confidentiality between them. With workgroups such confidentiality is maintained to the extent that members of one workgroup are not even aware of who the members of other workgroups are. It is of course possible to have people common to more than workgroup that can transfer knowledge between the workgroups or move people between them.
The broad scenario here is:
The idea is evaluated in 'Proposal idea' and a proposal made for a manuscript.
A separate workgroup is established for manuscript planning and development. There may be one such separate workgroup for each manuscript.
Manuscript planning gathers the necessary authors into their workgroup to produce parts of the manuscript. Members of one workgroup need not be aware of who the members of the other workgroup are.
The editor sends out the evolving manuscript for comment.
New events can be dynamically defined as new situations arise. The general approach is that once an event is created a new rule is also created to send a message to a role within an effected workspace.
Figure 7: Types of Interfaces and Commands
Supporting tools must provide ways to meet KIWA requirements, in particular implement the metamodel and allow users to workspaces through a presentation interface.
The process followed here would be to:
Step 2 - decide on the workspaces for the workgroup,
Step 3 - organize each workspace.
If a proposal is found to be of interest, a review process is initiated. The setting up of this review process is delegated to the office-manager, Rebecca.
Rebecca can initiate a standard interface for this activity and customize it to the any particular needs of this proposal.
The main workspace is now set using the development interface illustrated
in Figure 8. This workspace includes a number of windows for the different
kinds of workspace objects. There is a window that contains background
information such as publications policy; another contains documents used
by publications management; a window that shows the actions in this workspace;
and still another with reference to discussions, one of which is shown
in Figure 6. The discussion is a top-level discussion during project selection
stage. There are also windows that display roles, participants and actions
these roles can take. The actions can be used to initiate new workspaces
if needed. The publication manager can direct the office-manager to set
up the necessary process when a decision is made to proceed with a proposal.
The workspace here is simpler given the well-defined activity that this role performs. It provides the end user with the information in the workspace, access to other users and ways of maintaining discussions.
The system is based on a relational database that stores the workspaces. The server program accepts command strings that can be issued from a variety of interfaces, thus providing the flexibility to extend the system. This includes the wizards, special interfaces and agents described in earlier sections.
LiveNet also supports the creation of milestones rather than a strongly predefined workflow to support the kind of transitions defined by a transition diagram. Any number of events can be defined in a workspace. Rules can then be defined dynamically to define how events in one activity are notified to another. The occurrence of such events results in messages sent to other workspaces or roles within a workspace.
Cothrel, J. and Williams, R.L. (1999): "On-line communities: helping them form and grow" Journal of Knowledge Management, Vol. 3, No. 1, 1999, pp. 54-60.
Dourish, P. (1998): "Using Metalevel Techniques in a Flexible Toolkit for CSCW Applications", ACM Transactions on Computer-Human Interaction, Vol. 5, No. 2, June 1998, pp. 109-155.
Furst, S., Blackburn, R. and Rosen, B. (1999): "Virtual team effectivess: a proposed research agenda", Information Systems Journal, Vol. 9, No. 4, p. 249-270.
Hawryszkiewycz, I.T. (1999): "Workspace Networks for Knowledge Sharing" Proceedings of the Fifth Autralian World Wide Web Congference, AusWeb99, Ballina, Australia, (ISBN 1 86384 455 4), April 1999, pp. 219-227.
Jones, C.T., Hesterly, W.S., and S.P. Borgatti (1997): A General Theory of Network Governance: Exchange Conditions and Social Mechanisms, Academy of Management Review, Vol. 22, No. 4, October 1997, pp. 911-945.
Marketin, M. and Beranek, P. (1999): "Training to improve Virtual Team Communication", Information Systems Journal, Vol. 9, No. 4, October, 1999, pp. 271-289.
McAdam, R. and McCreedy, S. (1999): "A critical review of knowledge management models" The Learning Organization, Vol. 6, No. 3, 1999, pp. 91-100.
Nonaka, I. (1994): "A Dynamic Theory of Organizational Knowledge Creation" Organization Science, Vol. 5, No. 1, February 1994, pp. 14-37.