This session presents a Semantic Web location model; an RDF map of physical spaces and the connections between them. The tools, including a REST-ful interface and Instant Message bot, allow groups of people to build and annotate collaborative maps, building up a corpus of open geographical data and providing a group filtering system for recommendations. The model can be viewed as a MUD-like environment or as an application for a loose network of services such as an 802.11b network.

A model 'mudlondon' has been developed online for some months. Various online resources are used to provide cross-references for different geodata co-ordinate systems.

The application has two complementary interfaces. There is a REST-ful RDF web interface which is designed to be crawlable by an RDF spider, provides a uniquely identifying URI per location, and is designed to deliver and accept RSS-style XML feeds from other sites such as http://grault.net/grubstreet/ , the 'Open-Source Guide to London' wiki.

The primary user interface is an Instant Message 'bot' that uses Jabber, the XML-based messaging system offering 'transports' to most other IM systems. The bot presents the RDF model to the user like a MUD or multi-user virtual world, where users can explore, augment and comment on it in conversation. The bot uses the REST interface as a 'brain'. User metadata is stored in the FOAF, Friend of a Friend namespace which provides a basis for filtering trust networks. The model is supported by an ontology expressed in DAML, the Darpa Agent Markup Language.