Knowledge Maturing in Collaborative Spaces

Highlights

• Collaboration spaces consist of mixtures of knowledge at different levels of maturity, and tool support needs to address the complexities associated with this mixture (which is, additionally, dynamic in nature). This can be achieved through providing cues for the respective maturity levels also to users of collaboration spaces who have not been part of the knowledge development process.

• Outdatedness and relevancy as a temporal element in collaborative spaces are key concepts that augment the view of maturing of collective knowledge and address a key scalability barrier when those spaces become larger in terms of number of artefacts, number of participants and usage time periods. It enhances the ability of users to make judgments about the suitability of the prospective resources.

Research Challenge

Collaboration spaces in which knowledge is co-created and developed using artefacts, such as documents, are highly dynamic. They contain emerging knowledge, controversial opinions, but also artefacts that represent broad agreement, or representations of knowledge that has been superseded by knowledge represented elsewhere. While a limited inner circle of active contributors might be able to cope with this mix of various levels of maturity of knowledge and its representation, this constitutes a major barrier to informal learning in larger networks, particularly to new and peripheral members of such collaboration spaces, but also when distinct sub-groups have formed, that is when the knowledge how to deal with the information in the collaboration space is no longer sufficient to assess whether a given resource can be relied upon, or whether it is controversial, work-in-progress, or even outdated.

Theoretical Contribution

The knowledge maturing phase model [1] describes collective knowledge development along a developmental process which can be divided into phases each of which has different characteristics with respect to the knowledge and how to work with it and its representation in artefacts.

Knowledge Maturing Phase Model from [1]

One gap in the technology support for the knowledge maturing process that LivingDocuments [2] attempts to fill is the following: While many collaborative editing solutions focus on the step from phase II (distribution in communities) to phase (III) transformation, they do not acknowledge that artefacts (i.e., documents) do not correspond 1:1 to knowledge that they attempt to represent. Rather, documents are usually mixtures of different phases of knowledge maturing, each of which with different requirements:

• Phase II concentrates on establishing a shared understanding and is opening up to new input. Controversial discussions are an integral part of this phase.

• Phase III, by contrast, concentrates on consolidation, on filtering the different viewpoints and coming up with a single agreed outcome.

This constitutes different conflicting requirements, which existing solutions try to cover by introducing annotation functionality. However, it is not sufficiently acknowledged that knowledge does not (only) develop through the co-creation of single documents, but that in real-world context this takes place in larger document spaces. In such document spaces, the number of contributors grows and does often not overlap between different documents. This becomes especially true in larger networks where it is the normal case that looking for a resource involves making judgments about resources created by others.

This introduces a new challenge to the process as the knowledge maturing phase is not easily recognizable by those who were not part of the conversation. For them, it could be phase II or phase III, even phase I if it constitutes, e.g., a write-up of an individual opinion. However, this is a major aspect in assessing the usefulness with respect to reliability for the situation in which a user wants to access a document. Here, additional cues for maturity of documents are required.

One key element in this is the phenomenon of knowledge obsolescence. As discussions in co-design processes have shown, which were confirmed within the knowledge maturing indicator experiment, this is too often a mixture of two distinct phenomena and a related third one:

• Knowledge becomes irrelevant in a specific context.

• Knowledge becomes invalid with new evidence.

• Knowledge is forgotten.

This introduces the new notion of relevance as an important influence for awareness in a community. Changing relevance does not change maturity, but it does change the focus of a collective. Relevance is here not just the limited perspective of information retrieval, but encompasses also what is considered state-of-the-art in a professional community, which in turn is influenced by values (example: ecological construction), i.e., the identity of the collective (be it a network, company, community of practice or similar). Especially off-mainstream, knowledge strands are rediscovered that are given relevance again, and the assessment of relevance of these knowledge strands is one defining element to be a distinct collective. With diminishing attention, such “irrelevant” knowledge no longer develops. But also other strands of knowledge development get influenced; it becomes less connected to other areas, e.g., the knowledge how to apply a certain technique for a certain problem evolves so that it includes other techniques. The less connected it becomes, the more likely [3] it is to eventually disappear from the collective. As individuals are in multiple collectives, this could be a path for reintroducing this knowledge into a collective.

The phenomenon of knowledge becoming invalidated can be best described by a back-loop from a higher maturity phase to a lower maturity phase and a merging with the strand of knowledge development that has yielded the new evidence. The more disruptive the new evidence is, the larger the loop can become, especially it might involve discussion processes to reach agreement over the implications of the new evidence. But there is still merging going on as the new strand contrasts to the old.

Scenario

Clinical guidelines such as the NICE-Guidelines created on a national level need to be contextualised and implemented at a local practice. This is done with the help of devising an action plan that encompasses the shared understanding of how to apply the national guidelines in local implementation. This needs the involvement of all practice members to take into account their own experiences and to ensure awareness and adoption of changes to these plans. Previously, the action plan was mainly developed in a top-down way, using traditional documents. But the adoption of guideline changes was rather slow and unsystematic. This was related to lack of awareness about changes.

The action plan is collaboratively developed as an online document using LivingDocuments. Usually, the moderator of the process will start by creating a document, which represents an action plan for a specific guideline. He will then attach the national NICE-Guidelines to the document, invite his colleagues to edit, comment and develop the document further. The working group can add at any time - even if there is currently no active planning process - their experiences, which can be reviewed as soon as there is a revision of the action plan. The moderator monitors if there is consensus or disagreement. In case of consensus, the document is considered to be in a stable version and is then exported to any of the required formats like PDF or MS Word. All the concerned colleagues in the Practice are then notified about the new implementation plan. The plan can be finally discussed in one of the Practice monthly meetings before it is fully put into practice.

Findings and Design patterns

As part of co-design processes the following set of technical features have been identified as key design elements (see section on LivingDocuments for a more detailed description of the respective patterns):

• Collaborative editing with stable and unstable areas of collaboration. The two areas are represented by (i) a real-time collaborative editor, which supports both synchronous and asynchronous modes of collaboration, and (ii) an area for commenting on the document in which opinions can be expressed quickly and conversations can take place. Users can summarize the agreement reached in the commenting part in the stable document editor.

• Discussion and commenting. As conversations about specific issues can evolve into long and complex interactions, possibilities for discussions are offered that have an anchor in the document they emerge from. They take the form of sub documents in which again stable and unstable parts are available. This allows for mixing different levels of maturity within one document and accounts for the fact that “documents” usually cover more than one knowledge maturing process instance.

• Maturity awareness. For individuals who have not been part of the creation process, it is not easy to see to which stage progress has been made. Furthermore, as usually there is no systematic gardening of document spaces, it is not easy to recognize how the relevancy has been affected by external changes. This all makes it difficult to judge maturity of knowledge, which is needed to assess relevancy for a specific information seeking situation at hand. Furthermore, trusting knowledge contained in documents by unknown others requires cues to build trust relationships in the first place. This trust building can be supported through maturity cues as well as additional information about the authors of artefacts/artefact contributions.

• Outdatedness and relevancy as a temporal element in collaborative spaces becomes important as soon as (i) the size of the collaborative space increases, (ii) the duration of the collaboration extends over longer periods of time where different artifacts of the collaboration have different temporal characteristics, and (iii) the number of collaborating individuals increases, particularly when groups of creators and consumers of artefacts vary and overlap with respect to different artefacts. Computed indicators for these aspects need to take into account domain specific characteristics of artefacts.

• Awareness of knowledge maturing builds upon collaborative knowledge development practices- Knowledge maturing support is a socio-technical solution in which collaborative knowledge creation practices need to co-evolve with the respective tool support. Particularly cross-organizational knowledge development requires not only a change to collaboration practices, it also requires a change to prioritization of such activities in everyday work practices. This is a longer term change process.

Methods & Evidence

• Co-design process for LivingDocuments

• Knowledge Maturing Indicator Experiment in which a still active collaboration space of a collaborative research (after two years, with still two years to go) project with over 500 documents was tagged by project members according to outdatedness and relevancy. Subsequently, a machine learning algorithm was applied on the editing history and the content. Results suggest that certain features are good predictors for outdatedness and relevancy, but collaboration practices that involve various collaboration tools make it difficult to capture the maturing process. The more a single collaboration space can offer all the functionality, the better results will be. As the collaboration space itself and the data used was similar to the LivingDocuments system, the results appear to be transferrable.

Links

• LivingDocuments

• Knowledge Maturing (https://knowledge-maturing.com)

Reflections

Experiences and findings seem to suggest the following practical implications:

• Both relevance and invalidity require negotiations in the collective that often result in an agreement which of the two cases applies. This reinforces the need for solutions such as LivingDocuments that provide negotiation spaces that are contextualized and tie together different levels of maturity: Negotiations involve conversations around often more mature knowledge and their artefact representations [3].

• While invalidity has a more universal character and is not expected to change over time, relevance is clearly highly contextual and might change over time. This needs to be reflected in the way learners access, e.g., artefacts that represent such knowledge [4]. What is considered “invalid” is hardly interesting but for historic purpose, while access to what is considered “irrelevant” in “normal” contexts, becomes relevant in specific contexts, such as maintenance for old machines (more phase IV+) or even for emergent ideas (phase I). Indicators would be helpful for distinguishing the two cases as they will be treated differently for search and notification purposes.

Contributing Authors

• Martin Bachl (HsKA)
• Zahra Grosser (HsKA)
• Christine Kunzmann (PONT)
• Andreas Schmidt (HsKA)
• David Zaki (HsKA)

References

1. R. Maier and A. Schmidt, “Explaining organizational knowledge creation with a knowledge maturing model,” Knowledge Management Research & Practice, vol. 2014, no. 1, pp. 1–20, 2014 [Online]. Available at: http://www.palgrave-journals.com/kmrp/journal/vaop/ncurrent/full/kmrp201356a.html
2. M. Bachl, D. Zaki, A. Schmidt, and C. Kunzmann, “Living Documents as a Collaboration and Knowledge Maturing Platform,” in International Conference on Knowledge Management (I-KNOW 2014), 2014 [Online]. Available at: http://publications.andreas.schmidt.name/2014_IKNOW_demo_livingdocs.pdf
3. C. Kunzmann, A. P. Schmidt, and C. Wolf, “Facilitating maturing of socio-technical patterns through social learning approaches,” in Proceedings of I-KNOW 2015, 2015 [Online]. Available at: http://publications.andreas.schmidt.name/IKNOW2015_Facilitating_Maturing_of_Socio_Patterns_a.pdf
4. A. Schmidt and C. Kunzmann, “Designing for knowledge maturing: from knowledge-driven software to supporting the facilitation of knowledge development,” in International Conference on Knowledge Management (I-KNOW 2014), 2014 [Online]. Available at: http://publications.andreas.schmidt.name/2014_IKNOW_schmidt-kunzmann.pdf