RE: article on URIs, is this material that can be used by the SWEO IG?

Tim Berners-Lee:

| There are two ways of identifying things in generality, 303 and #,  
| and the TAG and arch doc discuss those.  Don't hold the TAG  
| accountable for the Topic Maps system of subject identifiers, or  

On a side note, in fact the 303 solution is not that different from subject
identifiers. Both assert whether a URI refers to an information resource or
not. In the XML serialization of Topic Maps, if a URI refers to an
information resource it is embedded in a <resourceRef> element, otherwise in
a <subjectIndicatorRef> element. So, tweaking the Topic Map definitions a
bit, it would be fair to say a 303 response means the dereferenced URI is a
subject indicator (the tweaking is 303 says "The associated resource might
or might not be an information resource" whereas Topic Maps differ whether a
subject indicator could refer to an information resource or not, depending
on which Topic Map standard is followed). The main point of subject
indicators is not the fact that some web page is 'about' some resource - the
main point is that someone has decided to use this URI as an _identifier_
for a resource.

The difference in the solutions is Topic Map subject indicators allow us to
use http://www.microsoft.com as an identifier for Microsoft Corp., and the
303 solution only allows this if we can convince MS to return a 303 from
their home page - which is either unlikely or would take time. I personally
consider this a good thing in the 303 solution, since hijacking existing
URI's for ontologies is vulnerable - if the Wikipedia editors decided to
change http://en.wikipedia.org/wiki/Shakespeare to
http://en.wikipedia.org/wiki/TheBardOfAvon, what would that mean for the
identification of him? 303 at least requires ontology makers to own the URI,
and hopefully guard its consistence over time.

Marc de Graauw

www.marcdegraauw.com

Received on Tuesday, 5 June 2007 22:11:33 UTC