Krzysztof Janowicz
What prevents the Geospatial Web from taking off is not a missing architecture and protocol stack but, beside other aspects, the question of how Web services can be semi-automatically discovered and whether and to what degree they satisfy user requirements. Two approaches turned out to be useful for semantic-enabled geospatial information retrieval: subsumption reasoning and similarity measurement. However, while the former one can be applied to query service ontologies described in OWL-S or WSMO/WSML, most existing similarity theories are not able to cope with logic-based service descriptions. This chapter presents initial results in developing a directed and context-aware similarity measure that compares WSML concept descriptions for overlap and therefore supports retrieval within the upcoming Geospatial Web.