Date: Tue, 12 Sep 2006 01:47:18 -0500
From: Sol Swords
CC: Dan Connolly
Subject: Wednesday's talk with Dan Connolly

Hi all,

We'll be meeting this week as usual, Wednesday at 4 PM in ACES 3.116.  
Our speaker will be Dan Connolly of w3.org; his abstract follows.

 - Sol


The Semantic Web is an attempt to integrate data across applications and 
administrative boundaries and at large and small scales. It is intended 
to do for data what the original Web technologies did for documents.

A central tension in knowledge representation is between expressiveness 
and efficiency: the more expressive a language is, the more 
computationally expensive it is to reason about.

Techniques for maintaining efficiency are akin to techniques for 
maintaining link integrity in hypertext systems: they are useful and 
necessary on a limited scale, but must be relaxed at the global scale.

On the Semantic Web, many different engines will produce proofs for 
different sorts of problems, but no one engine is expected to be 
complete for the whole set of problems. Parties are expected to be 
complete in their ability to follow a proof generated by any other party 
and judge whether it is valid.

Parties that don't use logic to justify their data can use digital 
signatures and logical reflection techniques integrate signatures into 
the proof system so that trust policies can be manipulated in-band with 
the rest of the data.

I will briefly review the Semantic Web technologies that have been 
standardized (RDF and OWL) and discuss current research work on 
representing proofs in the Semantic Web and using proofs to enforce and 
audit access and usage policies.