Collective Cognition with Semantic Mediawiki: Lessons and Experiences

Web-based tools have fundamentally revolutionized the ways of knowledge acquisition, propagation, aggregation, understanding, and analysis. This is exemplified by Wikipedia, one of the most successful Web tools that support collective knowledge management in a large user community. Semantic wikis are extensions to wikis with semantic technologies that are aimed at further facilitating better human understanding and automated knowledge processing. Common features of semantic wikis include the incorporation of semantic markup into wiki pages which can be further translated into Semantic Web languages, e.g., RDF or OWL, and querying mechanism for better knowledge retrieval and propagation. Therefore, semantic wikis are capable of supporting two kinds of knowledge representations: informal knowledge represented as natural language text (as normal wikis support) for human reading, and formal knowledge representation with explicitly defined semantics that is primarily for automated machine processing. Among all semantic wiki systems, Semantic Mediawiki (SMW) is probably the most successful one. We therefore use SMW as the representative of semantic wikis for our study and evaluation.

To reproduce the success of Wikipedia, a semantic wiki needs to follow some proven working principles, such as collaboration and ease of use. However, our experiences with several real-world projects reveal that semantic modeling is substantially more challenging than writing a conventional wiki for common users. The goal of this paper is to summarize some of our observations on key issues in collective cognition using SMW, including the choice of knowledge modeling patterns, the context and organization of knowledge, and collaboration protocols.

View Publication

Citation