Gazetteers and Geographical Information Retrieval

Gazetteers are coming to play an increasingly important role in geographical information retrieval on the web. They enable users of transport timetables, routefinders, yellow pages, web mapping services and geographical web search engines to employ place names when specifying the geographical context of their requirements. The use of gazetteers in this role has also served to highlight some of their limitations with regard to the needs of the user. In practice, many queries that specify place names fail. One of the prime reasons for this is that the user may employ an informal or vernacular place name that is in common use but which is not recorded in the available gazetteers. In the UK, examples of such names are the “Midlands ” the “Chilterns ” and the “Wye Valley”. The reason the name would not be recorded is that gazetteers tend to reflect an administrative view of the world with an emphasis upon places that have precise boundaries. Some gazetteers do record the names of topographic features such as mountains and valleys, but they are not usually accompanied by data that record an estimate of the spatial extent of the features. The existing gazetteers may also fail to recognise a name because they lack the required level of detail or geographical extent or simply because they are out of date. There is a need therefore for richer gazetteers that reflect common knowledge of place names. Because of the high rate at which place names change or are introduced there is also a need to develop a system of interoperable web gazetteer services that reflect local and regional knowledge of places throughout the world. For the purposes of geographical information retrieval it is possible to envisage an ideal situation in which there is a system of multilingual gazetteer services in which the content conforms to agreed methods for specifying: preferred and alternative names; the timeframe for use of names; an ontology of place categories; rich information on spatial context including geo-political and topographic hierarchies, coordinates in well defined reference systems, spatial relations to adjacent places, and spatial footprints at different levels of generalisation, with information on the nature of boundaries (precise / vague)