- From: Tim Berners-Lee <timbl@w3.org>
- Date: Mon, 15 Oct 2007 16:07:21 -0400
- To: W3C-TAG Group WG <www-tag@w3.org>
- Cc: Alan Ruttenberg <alanruttenberg@gmail.com>, Jonathan A Rees <jar@mumble.net>, Dan Connolly <connolly@w3.org>
At the Cambridge Semantic Web Gathering a few days ago I was was chatting with Jonathan Rees and Alan Ruttenberg from Science commons about basically web architecture from the semantic web point of view. Alan felt an urgent need for much more concrete basis for this than he could get by trying to red the current AWWW with semantic web- colored glasses. He, as , would really like to have an ontology for the things the AWWW document talks about, and a formal definition of the semantics of things like HTTP fetches, hash, etc. At the same time, Anne van Kesteren has been suggesting that the HTTP spec doesn't have a very clear semantics. He asks, for example, what happens if a server sends two different content-type headers, for example? There are no HTTP validators, ad the significance of it is not obvious. This problem could also be helped partially by some semantics expressed more explicitly. Two question Alan asked recently (on the list and offline) were - "How can one ever show that a web site is behaving contrary to the web architecture?" and - "How do i know what triples an RDF system is able to draw from an HTTP interaction?" Both god questions. The answer to the first question could be to draw all the triples from the HTTP transactions and the documents published, and then check for OWL inconsistencies. Which begs the second question. This is relevant to the Tabulator project, as Tabulator does this, and uses the conclusions from HTTP transactions to (for example) select user interface operations to offer the user, and to generate warning messages about inappropriate behavior. We wondered whether it would be good idea to put together some kind of a task force under the TAG to propose set of these axioms and an ontology. Tim PS:
Received on Monday, 15 October 2007 20:07:51 UTC