Semantic Web Applications at W3C

Some ideas for semantic web services for the W3C web site.

Access Control

W3C currently uses RDF to represent access control for resources on the www.w3.org site. The primary store for the RDF data is actually a database, but W3C staff are working on making the RDF representation in files be the primary store to facilitate automatic replication.

Versioning

W3C maintains document histories in a CVS repository. Making the older versions available via the web is one project, but making the history relationships available via RDF is the relevant project here.

Searching and Navigating

We wish to make it easier for W3C team, members, and the public to find information at www.w3.org. There is lots of information available about lots of subjects, but it is not easy to find just what you want amidst all the less relevant material.

Searching as we know it today could be improved, but only with considerable work of the kind done by search services. This would be probably be fairly expensive, and still amount to a large portion of guess work. Generally, searching can be done better if each page has metadata associated with it that would be useful for searchers to find the page.

We also want to improve the navigability of our site for browsing starting from any known page to find relevant material on other pages linked from the first. Supporting navigation well usually translates to providing a well-organized space presented in a way that is easily understood by users. This also takes a lot of work both initially, and over time as the organization changes.

Support for both searching and browsing can be facilitated by associating metadata with every page. The metadata would be indexed for use by search engines. We could also include a menu of links from every page to other closely related pages, and indexes of related pages.

A navigation bar could be included in every page, either manually (though this is error prone and harder to maintain) or automatically by a page generator service, or perhaps by using a standard stylesheet.

The navigation bar should reference a few standard things in predictable places. Which things are listed should depend on what kind of page it is. A tech report, for example, should list the main tech report page, the activity, the working group, the authors, and the W3C web site. We already have some of these things linked in the bars along the top of the page. But more could be done.

Moreover, the information that could be used to automatically generate such navigation bars, as well as provide searching support, should be made available in a standard form. The form we suggest is RDF because it provides a uniform way of relating resources to other resources.

The RDF info associated with any page should be available as RDF to those who should have access to it.

Some RDF info could be inherited from collections rather than being repeated for every page. Inheritance is not only a space savings, but moreover, it also allows for automatic update of some data in all instances that inherit the data.


Daniel LaLiberte liberte@w3.org
Last modified: Thu Dec 9 09:59:56 EST 1999