To propose methods for expressing semantics and operating semantics in largely distributed environment, such as peer‐to‐peer (P2P) based digital libraries (DLs) where heterogeneous schemas may exist and the relationships among them must be explicated for better performance in information searching.
In conventional solutions, a mediator is adopted to create and maintain the matching between relevant terms such that distinct but relevant metadata schemas can be integrated according to the mapping relationships in the mediator. However, such solutions suffer some problems originated from the static matching in mediator. This paper proposes to use facts to express the relationships among heterogeneous schemas and conduct the reasoning dynamically by using inference engines.
It is justified to use facts and inference engines to express and operate the semantics among heterogeneous but relevant information resources. The user can choose to convert only part of the XML document into facts if she can unpeel deeply nested XML tags. Additionally, it is possible for the user to manually edit (assert, update or retract) the facts as needed in the reasoning.
The study assumes that peers are clustered according to shared topics or interest. An exhaust evaluation has not been conducted.
Each node can publish its schema to the involved peer community such that other peers can automatically discover the specific schema. A local matchmaking engine is adopted as well in order to automatically generate the relations between its own schema and the retrieved ones.
This paper provides a framework for semantic data integration in P2P networks.
