> -----Original Message-----
> From: ext Aaron Swartz [mailto:[log in to unmask]]
> Sent: 30 October, 2001 16:43
> To: [log in to unmask]
> Subject: Re: [POLL] What is at the end of the namespace?
>
>
> On Tuesday, October 30, 2001, at 07:06 AM,
> [log in to unmask] wrote:
>
> > What is at the end of a namespace? Nothing. Nada. Zip.
> >
> > Namespaces are punctuation. Nothing more.
>
> While this may be true if you only follow the spec, there is a
> widespread expectation that namespaces should have something at
> their end (meaning here that if you paste an HTTP one into your
> web browser it should work). The W3C promotes this and does work
> that depends on this. Many users similarly like to visit a
> namespace to see what it's about.
>
> It's a good practice, IMO, and I see no reason to stop.
Sorry, it might be understandable for folks who see a URL used
as a namespace URI to expect to find something "at the end of it",
but (a) that means that the user doesn't understand the role or
purpose of the namespace URI and should be *educated*, and (b) if
the W3C encourages such continued misunderstanding, that is IMO
irresponsible and in fact a little puzzling.
I guess we should just stop talking about URIs or URNs at all
and insist that everything be a URL and that everything resolve
to something. But of course, that still doesn't address the fact
that a namespace does not denote a content model, so the example
of e.g. the XHTML namespace resolving to information on the
XHTML content model is very dangerous -- as I might define a
content model that draws on constructs/terms from the XHTML
vocabulary and imposes different constraints or interpretations
(considered compatible with the semantics defined by the XHTML
specifications) and yet does not conform to any of the XHTML
content models -- so just how does any application which obtains
information (schemata, etc.) from the namespace URI of a given
XHTML named construct actually know that it provides correct and
relevant information in the context of *my* usage. It can't.
And, by the way, just which content model does the namespace equate
to anyway? Strict? Transitional? Eh?
Again, a namespace is punctuation. To presume, suggest, or insist
that it is anything else is asking for trouble and taking a very
dangerous narrow view of the long-term big picture of how
automization can be realized on the web, semantic or otherwise.
Yes, it might work for clever toy systems, but it is not a reasonable
foundation for defining the automized interchange of knowledge
on the web.
Since there are so many folks already abusing URLs as URNs, why
the heck not also abuse URLs as resolvable namespace URIs? Bad
practice follows bad practice...
I think the W3C is great and has accomplished great things, but
that doesn't mean that TimBL is divinely infallible or that
any particular view or practice "blessed" by the W3C is right.
Simply saying "well, the W3C does it that way" is not a valid
argument. That's religion, not computer science. If the arguments
*why* the W3C follows a given way of doing things are valid,
fine, but one has to make those arguments stand on their own, not
based soley on who makes them.
> If you
> want to argue that we should, take it to some W3C list, not the
> Dublin Core itself.
Uhhh... with all due respect Aaron, the question was *asked*
here, not on a W3C list -- and yes, I've already brought up the
topic in W3C space, and will continue to do so.
And, no, I'm not interested in arguing at all. Open discussions
are always great, and are the way to move forward. But the key
word here is "open", and politics and religion shouldn't get in
the way, eh?
Cheers,
Patrick
--
Patrick Stickler Phone: +358 50 483 9453
Senior Research Scientist Fax: +358 7180 35409
Nokia Research Center Email: [log in to unmask]
|