W3C home > Mailing lists > Public > xml-uri@w3.org > May 2000

Re: A little courtesy, please

From: Dan Connolly <connolly@w3.org>
Date: Tue, 23 May 2000 09:58:23 -0500
Message-ID: <392A9C8F.6C040C53@w3.org>
To: "Simon St.Laurent" <simonstl@simonstl.com>
CC: Tim Bray <tbray@textuality.com>, xml-uri@w3.org
"Simon St.Laurent" wrote:
> At 03:05 AM 5/23/00 -0500, Dan Connolly wrote:
> >> [* re RDF] - the notion that sticking to what the namespace rec says
> >> tends to destroy RDF is just totally 100% incontrovertably vacuous.
> >> RDF has chosen to say that when you're in RDF space, namespace names
> >> have to be used in a particular way.  Some of us may have trouble with
> >> the syntax engineering (#, feh) and some of us may worry that the RDF
> >> way doesn't provide enough indirection, but nothing they do contravenes
> >> the namespace spec in spirit or in letter.
> >
> >I largely agree with this, but I cannot agree that treating relative
> >URI references as namespace names without absolutizing them is in
> >the spirit of RDF.
> This is, at least in theory, a discussion of allowable XML syntax, not RDF.
> I've seen no evidence that RDF processors are incapable of handling
> absolutizing within their own layer of processing.

Yes, to the letter of the specs as drafted, this is the case.
His claim went to the sprit of RDF, and from what
I can tell, the spirit of RDF is that xmlns works like href,
where relative URI references aren't intended to be compared
across documents without expanding them to absolute form.

> > Either namespaces are web resources in every sense of the word,
> > and hence any sort of URI reference the author chooses
> > may be used to point to them, or not.
> You're overstating your case here.  Namespaces use URIs as names,

Would that they did! The problem is that the spec says they
use URI references for names:

	"[Definition:] The attribute's value, a URI reference,
	is the namespace name [...]"
	-- http://www.w3.org/TR/1999/REC-xml-names-19990114/

In conventional usage, a URI reference denotes a absolute URI with
respect to some base URI, and the absolute URI denotes a resource.
But in the namespace spec as presently written, a URI reference
denotes a namespace directly, without going thru
the "denotes a URI" step. Hence this discussion.

> and
> nothing more, at least at the XML level, as described in the Namespaces in
> XML spec.
> URIs are not Platonic Forms, with austere and unchanging nature - they're
> useful tools treated differently in a variety of contexts.


Resources on the other hand, _are_ abstract things
like platonic forms...

> >And if not, the design
> >of XML namespaces doesn't agree with web architecture, which
> >is that important things should be treated as web resources
> >with all the rights and obligations thereof.
> At which point my earlier (unanswered) concerns about whether 'Web
> architecture' and the 'Semantic Web' are actually a liability for XML come
> into play, and the courtesy level likely drops another few notches.
> This isn't a discussion of religion or morals, or at least didn't claim to
> be.

As I say, I'm also much more comfortable discussing black-and-white
test cases. But claims like

	"URI's simply do not have the ability to identify resources.
	They locate them."

	-- Fri, 19 May 2000 00:34:38 -0600 (MDT)

are not turing-decideable, and they conflict with the
sense of design that I believe is responsible for
the success of the Web.

During the development of the Namespaces spec, I did everything
I could to leave such claims completely alone, since they're
not directly relevant to running code.

But it's clear now that belief in such claims has real consequences:
404s when I go looking for namespace documentation, DOM specs
without support for base URIs, and so forth.

So I think that addressing these philosophical issues directly
is cost-effective at this point.

> While I'm glad to have been able to participate in the discussion, I'm
> really left wondering why this issue moved onto a public list when the 'Web
> architecture' that appears to be motivating it is under wraps, apparently
> unquestionable.

Web architecture is largely a collection of axioms, and axioms are
just that... "we hold these things to be self-evident..."
sorts of things. So I'm very interested in any questions you may
have, but I won't be able to answer them with logical arguments
where principle #7 of Web Architecture is the conclusion.
But I should be able to answer ala "if you don't follow principle
#7, you'll be inconsistent with deployed code that does XYZ..."
or "we've seen people try to go around principle #4, and
the result was a mess...". But sometimes the answers are
likely to be less satisfactory ala "we didn't predict
web crawlers and global search engines, but principles
#2, #3, and #4 were keys to allowing that to happen, and
we want to keep the future open to that sort of

Dan Connolly, W3C http://www.w3.org/People/Connolly/
Received on Tuesday, 23 May 2000 10:58:42 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 23:13:58 UTC