Binding Points for Subject Identity: The case for standard Published Subject Indicators

Exchanging information about a subject with semantic tools, and singularly topic maps, demands both human and computer readable ways of agreement about this subject identity. Since subjects are always addressed through representations, hence agreement about subject identity must be grounded on some sort of representation, be it a name, an addressable resource, an URL or an XML element in a document. An overview is made of how those different representations can achieve or not the task of indicating subject identity in a non-ambiguous way. To address this problem, the XTM specification recommends that subject identity should be defined as far as possible by reference to Published Subject Indicators. But requirements for any standard definition, structure, management and validation of those PSIs remain to be defined, and in fact such requirements were deliberately considered out of the specification scope by its authors. Propositions are made for such requirements, grounded on the claim that a PSI should be a binding point for every possible representation of