Determining HTTP URIRef Alias Equivalence for Information Resources

I'd like some comments on the feasability of this idea.

-- The Problem --
Although URIRef aliases are not harmful to graphs (unlike collisions [1]), they
make it hard to combine graphs.  For example, if my URIRef for AWWW is:

<http://www.pitt.edu/~jfcst24/2005/06/28/AWWW>

(note:resource above does not exist.... just an example) and the W3C uses:

<http://www.w3.org/TR/webarch/>

Then finding out everything I and the w3c assert about AWWW requires one
knowing both resource names.  This is a social problem and not a technical
problem, I think.  I've seen these kind of situations occur in ontologies on
the web [2] already.

-- Partial Solution --
For the specific case when talking about information resources accessed over
HTTP, I see some possibilities for alias inferencing without OWL.

* Say one URL redirects to the other URL.  In that case, then it is a good bet
they indicate the same resource.

* Say you download one representation from one URL via a 2xx code.  Then
download from the second URL a representation in the same media type as the
first URL.  If the media types and representations are the same, then the two
are probably the same information resource.

Those aren't a complete solution, but does it partially help determine URIRef
equivalance in some circumstances without having to use OWL?  Are there any
other ways of automatically determining URIRef aliases without using OWL (or
DAML)?

[1] http://www.xml.com/pub/a/2005/01/19/review.html on "URI Collisions"

[2] E.G. using

<http://norman.walsh.name/knows/who#danny-ayers>

or 

<http://dannyayers.com/misc/foaf/foaf.rdf>

(OK, technically DA is a blank node in that document but still...)



--
Jimmy Cerra
https://nemo.dev.java.net


		
____________________________________________________ 
Yahoo! Sports 
Rekindle the Rivalries. Sign up for Fantasy Football 
http://football.fantasysports.yahoo.com

Received on Tuesday, 28 June 2005 22:50:01 UTC