Future WWW Resource Discovery by Metadata Agents

The World-Wide Web is a rapidly evolving information-rich environment. In its present form, it provides user access only to those resources which are explicitly known. Solving the problem of resource location is crucial to users who wish to exploit the Web's resource base, and therefore provides a commercial opportunity based on resource discovery systems and services. Existing approaches to the resource discovery problem are based on Web robots and spiders which are used to build massive indexes of resources. This may be termed "just-in-case" resource tracking. These approaches are inflexible and do not scale well. The proposed new technologies based on metadata objects and mobile agents will be able to solve all of these problems. Metadata is needed so that the appropriateness of resources with respect to specified criteria can be determined without requiring access to the resources themselves. Searching will be carried out in a "just-in-time" fashion, search criteria will be encapsulated as metadata objects carried by mobile agents. These agents will conduct their searches based on user-specified criteria, expressed as scripts, to ensure the appropriateness of the resources returned. This presentation summarises the primary issues for future WWW resource discovery schemes, describes the infrastructure that must be built to support those schemes, and gives a snapshot of ANSA's involvement in the both development of appropriate WWW standards and the prototyping of these ideas.