Skip to main content

C-WEB: A Generic Platform Supporting Community-Webs

Exploitable results

In one year, the following results have been reached. A. Assessing users’ needs and survey of existing technologies: -We have worked with a small sample of representatives of large companies, plus one cultural organisation, to understand their current practices, technical choices, and difficulties related to document management and deployment of corporate information system. -We have, in parallel, looked carefully at the current technological offer for EIPs, search engines, semantic indexing engines, document management and knowledge management systems. The conclusion of this market survey is that the "semantic web" is still to be implemented, and that C-Web is a timely R&D project. B. System positioning: -Taking into consideration some real users’ needs, the existing technologies, and the kind of innovative services that our partnership was able to design, we have decided to position the future C-Web technology as a combination of a portal generator and of a mediator, with advanced document management features. C. System design: -To support the functionality which is briefly described in section II, the system is designed as a middleware, built upon a knowledge database. The knowledge database stores a C-Web schema, which results from the merging of an ontology and of various terminology resources. It also stores local metadata which are descriptions of the sources and of the resources participating in the C-Web. These descriptions make use of the terms (concepts and properties) defined in the C-Web schema. -The system imports and exports the C-Web schema, and the metadata, in the Resource Description Framework formalism standardised by the W3C. A schema editor will support users producing C-Web schemes. It will of course produce RDF statements. -Various types of resources and sources are described in the knowledge base, and therefore, can be retrieved through querying. Static XML resources can be published as collections when they conform to some DTD or XML schema. In such a case, the DTD or schema is published, i.e.: the relevant element types are described by conceptual properties. The XML instances can ultimately be automatically indexed, using the content of those elements that have been published. -Structured sources (e.g. relational database) can be published along a similar process; the semantics of predefined queries can be described using some terms and properties defined in the C-Web schema. In fact, any structured source is wrapped onto a XML structure, and seen from the C-Web mediator a wrapped source is similar to a collection of valid XML documents. Unstructured resources can be published “manually”, a user (author, producer, publisher or librarian) create the metadata describing such resources, using a metadata editor which will guide the user and enforce the use of the concept and terminology defined in the current C-Web schema. The knowledge-base can be queried through a query language implementing first and second-order query evaluation on the database. Various applications can be built upon this database system. The first target application is a portal generator, which automatically builds a portal by querying the knowledge-base. If some characteristics of the users are represented in the knowledge base, they can be taken into account in the querying process, hence enabling generation of specific portals for different classes of users. http://cweb.inria.fr