Re: REST as formalism of RFC 2396 was ... what the SW needs from the Web...

> Perhaps I am one of a small minority who believes fairly strongly that REST
> and RDF are compatible. In particular, and at somepoint soon, hopefully, RDF
> will learn to better deal with the need to tie the 'meaning' of a URI to
> some function of the resolution of the URI. At present, RDF can only make
> assertions about URIs but as TimBL says, the what the URIs themselves mean
> is determined by RFC 2396. If we accept REST as the theory behind RFC 2396
> then we need to integrate the REST meaning of a URI to the RDF assertions
> made about the URI. So the question ought not be frame as REST can or cannot
> do x, y, or z, but rather how can RDF make use of a REST determination of
> the meaning of a URI -- to paraphrase, how does the mapping of a URI to
> representation over time determine the meaning of a URI.

I don't agree with several assumptions you're making along the way,
but I'm happy with where you end up.  URIs are a particularly useful
class of strings because of certain conventions about how they are
used.  Understanding those conventions and establishing new
conventions to allow them to be used sensibly in RDF is both important
and possible.   But I'm still not sure how we'll get there from here.
Maybe a URI task force in the RDFIG?   Maybe something related to some
W3C URI activity; I dunno.

Something where the people involved in the discussion are actually
interested, willing, and fairly committed would help the tone of the
debate, I think.  The fact that most people on the TAG and www-tag
seem to be filled with fear and loathing at any hint of httpRange-14
does not really help the discussion proceed.

> You may notice that I can eliminate the term 'resource' by substituting the
> phrase 'meaning of a URI'.

Absolutely.  I find rephrasing to avoid using certain
potentially-ambiguous terms to be essential to this discussion.  I
started by avoiding the term 'resource', and I've been toying with
avoiding 'respresentation', 'identify', and URI-as-a-noun (preferring
instead "URI string" which helps remove imagined mystical powers of
URIs).
 
    - sandro

Received on Saturday, 1 February 2003 22:31:11 UTC