The Nature and Value of Geographic Information

If geographic information is to be traded as a commodity, it must be possible to measure its quantity, to compare the contents of two data sets, and to perform other simple functions such as the measurement of similarity of content. Shannon-Weaver information theory is found to be inadequate for this purpose. In current practice geographic information is priced by such medium-dependent measures as area, or data volume. A theoretical framework is proposed in which geographic information is defined by atomic tuples, to which other structures can be reduced. Although the potential number of such tuples is infinite, the ubiquity of spatial dependence and the impossibility of determining location exactly allow useful representations to be created from finite numbers of tuples. Based on this definition of geographic information, formal definitions are proposed for geographic information systems and geographic queries. It proves difficult to incorporate naïve geography within this framework because of its logical inconsistencies. A semantic theory of geographic information is proposed, based on networks of linkages expressed as tuples. Unresolved questions are identified.