Using Semantic Searching for Web Portal Interoperability

The Web can be viewed as an open heterogeneous information space indexed by search engines and directories based on keywords. The indexing techniques fall short of providing satisfactory efficiency/effectiveness and recall/precision parameters which are required by information search and retrieval systems. The use of metadata for resource discovery would improve these search and retrieval parameters. However, the incompatibility of formats for exchanging metadata, metadata spam or even lack of widespread knowledge about metadata has inhibited this evolution. To improve these parameters and obtain a higher level of performance for searching and retrieving over the Web, interoperability and trust-ability are some of the fundamental factors to be improved. To solve these problems some projects try to create an infrastructure for interoperability based on metadata to give more semantics and organization to the Web. However, there is almost no agreement about the metadata element sets to be used, they do not catalogue all the information and they do not enter into the places where the information really lives, the databases. This paper raises issues of semantics based on metadata, ontologies and Z39.50. Additionally, it presents an architecture for an interoperable and distributed infrastructure over Thematic Portals, providing searching and retrieval services from the Web.