The original workshop proposal covers the more general problem of
information integration on the web; for this workshop, we have (at the
suggestion of many reviewers) chosen to focus the work on service
integration on the web. Nevertheless, we think that the original proposal
has some interesting context information that could be helpful to
prospective participants, so we have included an excerpt here.
AAAI-02 Proposal for Workshop on Intelligent Integration of Information and Services on the WebThe plethora of documents and applications available on the World Wide Web today presents a great opportunity to users: we can now access huge amounts of information on any topic of interest, and we can use a range of services to perform a variety of every-day tasks. This availability also presents a great challenge - how to reuse all this readily available information and services to construct new information and to deliver new, more complex and more valuable services.Information originating from different sources needs to be integrated in order to gain full advantage of having that information available. For example, information retrieved from different news agencies could be integrated and summarized to provide a compact and comprehensive news briefing. Similarly, currently independent services can be combined to deliver more value to the customer. For example, given a set of applications offering a particular type of goods, a comparison shopping service could be provided, or given a set of applications providing parts of a complex artifact, a supply-chain management service could be built for just-in-time construction of this artifact. Today each different source publishes information in its own proprietary syntax and following its own semantics for the described entities, usually derived from the schema of the internal application database supporting the services of the source. XML is a step towards a solution to this problem since it provides a means for defining a standard syntax for a service domain; if two distinct applications published their information using the same schema, i.e., tags defined in the same namespace, then their data are in syntactic and semantic agreement and can be exchanged and integrated. XML schemas, however, do not have explicit semantics, so if two applications do not share the same schema their data are completely incompatible, even though they may actually share the same semantic view of the domain. To make matters worse, precious little information is published directly in XML. In most cases "legacy" sources use their own formats, using HTML or structured-text conventions, and integration implies the post-publication information translation to an XML notation based on a particular schema. The schema chosen for this step may be constructed for the particular source or it may be a generally agreed upon schema for the application domain. A further problem hindering the integration of services is that languages, such as WSDL and WSFL, designed to declaratively specify services and their integrations also lack semantics. Thus, reasoning about the properties of the integrations, including correctness against organizational policies and performance, is not possible. In this workshop, we plan to bring together researchers developing languages for making explicit the semantics of the information and services available on the web, methods for using these languages to reason about them, and architectures for supporting the design and run-time execution of integrated applications on the web. In particular, we invite researchers working in any of the following areas to participate:
|
AAAI