Semantic wikis in knowledge management for multidisciplinary product development.

Albers, Albert ; Sauter, Christian ; Meboldt, Mirko 等

Abstract: Distributed workplaces, increasing product complexity and rising numbers of involved persons and domains are only a few reasons why the development of multidisciplinary systems is a challenging task. A lot of knowledge is needed for and generated in development processes. Knowledge management is gaining importance as a success factor in product development. The application of semantic wikis can support product development processes as tools for documentation and retrieval of explicit knowledge as well as for distant asynchronous communication and opinion forming between development team members.

Key words: wiki, semantics, knowledge management, collaboration

1. INTRODUCTION

The development of complex products is becoming more and more knowledge intensive. The quality of the management of product and process knowledge is an increasingly important factor for successful and efficient development processes. Technology-centered approaches to store all information in one single information technology tool have failed so far. In this article we suggest to apply semantic wikis as knowledge management systems, which on the one hand are easy to use and flexible knowledge storages that contain some of the generated product and process knowledge and on the other hand work as "central knowledge nodes", which cross-link wiki contents with other wiki content and other knowledge sources outside the wiki. Semantic wikis are a relatively young technology that implements some of the concepts of the Semantic Web (Berners-Lee et al., 2001). The semantic features complement the usability of wikis with additional functionalities for structuring and retrieving the contents of a wiki.

2. PROBLEM

Today the entirety of knowledge in the development of a complex product can no longer be comprehended by one single person. Effective knowledge management for product development processes becomes increasingly important. So far three different generations of knowledge management approaches are being distinguished in literature. The first two phases were dominated by a strong belief in technical solutions for supporting decision making and including all relevant knowledge in one system. In the third generation, the focus has changed more towards the human being and knowledge management through interaction between humans as well as between humans and knowledge management systems. Wikis and semantic wikis are technologies that implement ideas of web 2.0 and the semantic web. These technologies on the one hand enhance (distant) interaction between humans via internet and on the other hand support documentation and exchange of explicit knowledge as an easy to use publishing tool.

3. WIKIS

Wikis are software systems, which allow users to easily generate, publish and edit web pages (Cunningham & Leuf, 2001). Wiki is a Hawaiian term and means "fast". The first Wiki was implemented by Ward Cunningham in order to easily exchange information in a software development project. His intention was to create "the simplest online database that could possibly work". Today user-friendliness is still a central goal of wiki development. Two main elements of a wiki are the wiki pages and the wiki engine. The wiki pages are created and edited by users and contain the actual content. The wiki engine is the software system, which provides the functionalities for viewing, editing and publishing the wiki pages on the internet. Wikis have become increasingly popular in recent years mainly due to the following advantages:

* Easy collaboration and formation of opinion

* Easy documentation and editing

* Easy cross-linking

* Simple structuring

* Full text search

* Open source

Wikis are currently being used in a wide range of applications. The most popular and best known one is the online encyclopedia Wikipedia. Various other groups and organizations use wikis for organization and communication purposes. In the professional sector, wikis are often used as collaboration tools in software engineering projects (Decker et al., 2005).

4. SEMANTIC WIKIS

Semantic wikis provide the same functionality as normal wikis complemented by some additional features. Semantic wikis offer the possibility to enhance the contents of a wiki with metadata (Volkl et al., 2006). This metadata can be interpreted by software and makes it much easier to retrieve relevant information. Most semantic wikis use two types of metadata: relations and annotations. Relations indicate a relationship between the contents of two wiki pages. In comparison to simple hyperlinks, that are already a basic feature in wikis and in nearly all other web pages, relations are typed links and they do not only state the existence of a relation but they also contain information about the type of relation that connects two wiki pages. Annotations have a specified data type and explicitly attach pieces of information to the subject of a wiki page, which are not described any further otherwise, e.g. numbers, dates or names. Adding metadata formalizes and structures the contents of the wiki and it becomes partially machine-interpretable. Export functionalities to standards for the semantic web like RDF or OWL allow the reuse of the information stated in the metadata also outside the wiki. Metadata can easily be included in the contents of a wiki page with a simple syntax and minimal additional effort for the contributor. In the viewing mode, this metadata is not visible in the normal text for the reader, so it does not disturb the readability of the text. One motivation for using a semantic wiki instead of a standard wiki and for accepting additional effort for adding the metadata is that knowledge, which is relevant for a certain topic in a certain context can be retrieved more easily by queries not only on the text but also on metadata. Semantic wikis allow simple or complex, logically connected queries. These can be formulated on a special page or the results of a query can be included seamlessly into a page. These queries are executed every time a page is opened, so the page is kept up-to- date automatically and represents the current content. Another advantage of the semantic functionality is that it is often completely additive, so all advantages of normal wikis persist. If a user wants to quickly document an idea he can make use of the wiki as a fast and efficient publishing or documentation tool. If the content has ripened over time, metadata can still be added later on.

5. PROCESS MANAGEMENT AND DOCUMENTATION

Development processes should be structured according to a common conceptual model (Albers & Meboldt 2006). This makes them more comparable, transparent and makes it easier to reuse process knowledge from previous projects. Processes need not be identical but they should be describable with a common vocabulary and syntax. This enhances the transfer of process knowledge from a previous process or process step to a current one. Normally process iterations are only guided by the results of development phases like project reports or prototypes. Only the result, not the path on which this goal was reached, is normally considered relevant and documented. Unfortunately a lot of knowledge about the process path is lost on the way. The decision bases and the alternatives which were rejected on the way are normally not documented. If it is possible to formulate a common meta-model for product development processes in a restricted environment, this model can then be used in combination with an efficient documentation tool to structure the knowledge and to provide it in a suitable way. If semantic wikis are used for documentation, contents described in a wiki page are assigned to a specific process element via metadata. In case of an iteration or reuse of information from a previous project, this allows users to navigate through the process more easily or to formulate a query more precisely by also specifying process aspects. The possibility to find relevant information not only by categories and key words, but also by affiliation to a certain process step, increases the chances to find the right information.

6. MULTIDISCIPLINARY PRODUCT DEVELOPMENT

During the development of a complex multidisciplinary product, large amounts of new product knowledge are generated, which cannot be fully comprehended by a single person. This is disadvantageous but it is not absolutely necessary for the daily work of a development engineer. For his daily work only a subset of the total knowledge is relevant and needed. It is often difficult to identify these relevant knowledge elements. Semantic wikis can be applied as a means to decrease this problem. In the metadata, which supplements a semantic wiki page, authors can include information about the context in which the information was generated regarding process, product and domain in which the contents was developed or for which domains the content is known or assumed to be relevant. In the development of complex multidisciplinary systems, cross-domain interrelations between subsystems often lead to time consuming mistakes and iterations. Stating known interrelations over domains in the metadata helps developers to become aware of cross-domain effects. If a design change in one domain is made, 'documented cross-domain relations can help to identify possible subsystems in other domains, which might be affected by a domain-specific change. A page, which describes the embodiment design of a mechanical subsystem of a mechatronical system and links to the relevant CAD-files and drawing files would then contain in its metadata the information that the content is associated to the domain mechanics and that there are for example relations to the domains electronics (design space, electric or magnetic fields ...) and controls (kinematics, inertia ...). If all pages of a wiki are consequently enriched in this way, it is easier to identify relevant information by combining text queries with queries on metadata. Two prerequisites need to be fulfilled for this approach to work. Firstly an ontology needs to be formulated, which allows users to state the metadata in a consistent way and secondly sufficient amount of initial wiki page content and metadata need to be entered before the wiki can provide any benefits. Therefore the introduction process is a crucial success factor. Enriching wiki page content with metadata also makes it easier to retrieve and reuse information from earlier projects for current problems. Full text search, namespaces and categories help users in normal wikis to find the content they are looking for. If a semantic wiki is used to document a project according to the concept described above, engineers do not only have the possibility to search by key words, they can also specify the product context and process context as well as the domain for which the information is needed. This increases the chances that they actually find the information they are really looking for. If similar problems need to be solved more than a few times, it would also be possible to automatically create a collection of "best-practice"--examples by including an appropriate semantic query in a special wiki page.

7. CONCLUSION

A concept was presented how semantic wikis can be applied to support knowledge management in product development. By adding metadata to the contents of a wiki and the possibility to formulate simple as well as logically connected queries on metadata and conventional contents, reuse is simplified. Stating domain affiliation and domain relevance in metadata enables the generation of domain specific views.

8. REFERENCES

Albers, A.; Meboldt, M. (2006), A New Approach in Product Development based on Systems Engineering and Systematic Problem Solving, Proceedings of Applied Engineering Design Workshop, AEDS, Pilsen, Chzech Republic, October 2006

Berners-Lee, T.; Hendler, J.; Lassila, O. (2001) The Semantic Web. Scientific American 5 2001, 285ff

Cunningham, C. & Leuf, B. (2001) The Wiki Way. Quick Collaboration on the Web, Addison Wesley, 2001

Decker, B.; Rech, J.; Ras, E.; Klein, B. & Hoecht, C. (2005) Self-organized Reuse of Software Engineering Knowledge supported by Semantic Wikis. Proceedings of the Workshop on Semantic Web Enabled Software Engineering (SWESE), Galway, Ireland, November 2005

Volkl, M.; Krotzsch, M.; Vrandecic, D.; Haller, H.; & Studer, R. (2006). Semantic Wikipedia, Proceedings of IW3C2, Edinburgh, Scotland, May 2006