Copy the page URI to the clipboard
Collins, Trevor; Mulholland, Paul and Watt, Stuart
(2001).
URL: http://www.ll.unimaas.nl/euro-cscl/Papers/28.doc
Abstract
Many communities exist that learn and share information either partly or wholly online. These (wholly or partially) on-line communities share messages, documents, and other artefacts that contain useful community knowledge. Members of the community learn through this sharing process, and the growing archive they create forms a valuable learning resource for existing and new members of the community. Two main kinds of approach exist to support community members in accessing resources. The first kind associates each communal artefact with a conceptual structure that represents its meaning. This approach requires high levels of maintenance, especially when the community resource grows at a fast rate. The second uses statistical and text analysis techniques to (semi) automatically derive semantics from the resource. There is increasing evidence that artefacts constructed and shared within a community follow genres revealed in the structure of the artefacts and the terminology used. These implicit genres used in the community are invaluable to members in constructing and interpreting artefacts, but existing tools that support members in locating and classifying resources make little or no use of genre. Our preliminary findings demonstrate the potential of genre-sensitive classification and retrieval tools.