Topic Maps are the international industry standard for semantic information integration. Appropriate means for Topic Map exchange are crucial for its success as integration technology. Topic Map exchange bases on the governing Subject Equality decision approach, the decision whether two Subject Proxies indicate identical Subjects. This paper discusses the 'absence of shared vocabularies' in the context of these decisions. Thereby, a differentiation between Referential and Structuralist Subject Equality decision approaches is introduced. All existing approaches to Topic Map exchange base on the TMDM. This implies a Referential Subject Equality decision approach and bound to a concrete Subject Map Disclosure (SMD) ontology and Subject Map (SM) vocabulary. This paper introduces a Structuralist Subject Equality decision approach which is called SIM. It allows the exchange of Topic Maps in the absence of a shared SM ontology and SM vocabulary.
The challenge in an exampleWithin a cooking peer-to-peer network remote peers exchange recipes documented as Topic Maps 1 . To collect information, peers send Topics which represent the Subjects of interest to remote peers. In the cooking network a Subject might be 'roasted lamb loin'. The remote peers check the availability of information about this Subject and respond with an according Topic Map Fragment. Afterwards, the requesting peer integrates all remote recipes about roasting lamb loins into its local recipe collection.This works fine if all peers made agreements about how to describe lamb cuts correctly. What happens if a remote peer uses the term lamb saddle instead? Or roasted lamb leg chops? The resulting meals are identical, but the requesting peers will never receive their recipes from distance. This shows that two critical points arise, if semantic agreements are not made by all peers logging into the network: How to request knowledge from remote peers if shared vocabularies are not available? How to integrate (merge) the received information into the local Topic Map?The solution proposed in this paper allows peers to interact in networks without having the overhead of centrally enforced vocabularies. Our solution detects 1 To avoid ambiguities all terminology concerning Topic Map Technologies is capitalised.