W3C home > Mailing lists > Public > www-tag@w3.org > January 2003

Re: yet another sidetrack on what a URI identifies

From: Dan Connolly <connolly@w3.org>
Date: Fri, 24 Jan 2003 10:56:08 -0600
To: Tim Berners-Lee <timbl@w3.org>
Cc: "Roy T. Fielding" <fielding@apache.org>, Sandro Hawke <sandro@w3.org>, www-tag@w3.org
Message-id: <1043427367.24068.57.camel@dirk.dm93.org>

On Thu, 2003-01-23 at 23:33, Tim Berners-Lee wrote:
> On Tuesday, Jan 14, 2003, at 19:58 US/Eastern, Roy T. Fielding wrote:
> > I know why your model breaks.  Try describing caching.
> Ok, Roy, so using my model or Sandro's, show  why when you
> try to describe caching it breaks.
> >   For real,
> > not in your imagination, but on paper and with appropriate definitions
> > for the metadata that is included in HTTP responses so that the
> > client has an interoperable clue as to what you are talking about.
> > Then try describing hierarchical proxies, gateways, etc.  And if you
> > are willing to continue work on that model, use it to describe the
> > semantics of a Web to GSM SMS gateway using POST.  I already did that
> > stuff seven years ago,
> If you have it in the form of a formal system, then I think
> a pointer to it would be very interesting.   I know Dan Connolly did
> some work in Larch,

Yes; at the May 2000 web conference, I presented

  Specifying Web Architecture with Larch

which traces my thoughts on formalizing the web from
discussions with Burchard and company circa 1995,
problems with that simple model, fixes to various problems,
culminating in a model of HTTP caching that, after discussion with TimBL
and others, I find quite satisfactory...


specifically, the bit about...
	% The Conditional GET Axiom, If_Modified_Since case
	% c.f.
%html Ari Luotonen and Kevin Altis<br />
%html <cite><a href='http://www1.cern.ch/PapersWWW94/luotonen.ps'>World-Wide Web Proxies</a></cite><br />
%html Proceedings of the 1st International WWW Conference, May 1994.

The larch traits that I use to do my thinking about
web architecture

cover a variety of issues, from URI syntax
to RDF model theory
to transcribing (and finding a small bug in)
Wadler's 1999 formalization of XPath/XSLT

Perhaps more evidence that TimBL and I are not just
talking is the URI syntax implementation we've worked
on, including some test cases that represent issues
with RFC2396. I'm not sure we've reported them all


Actually, timbl and I haven't quite reached
consensus on whether relative URI references
are allowed in some parts of the uripath API.

But our discussions are very much based on running
code, machine-checked formalisms, and real-world
applications... such as automation of the W3C
digital library
and managing my own personal travel itineraries

As TimBL mentioned, cwm implements the integration
of all this stuff... connecting KR to WWW. The
log:semantics primitive in cwm is a simpler model
than the 9711theory/HTTP model... but I think
it's a useful simplification, the way Newtonian
physics is a useful simplification of relativistic
physics. I hope to elaborate on that analogy
sometime between now and our Feb ftf...

> > so I expect a little more than talk at this
> > point from the W3C.

It's tricky to manage the balance between development
and "talk" here in www-tag. I'll thank you not to
jump to conclusions, assuming we're only talking,
and to trust that we're each doing our level best
to be constructive.

Dan Connolly, W3C http://www.w3.org/People/Connolly/
Received on Friday, 24 January 2003 11:57:40 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:32:36 UTC