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

RE: [Fwd: Re: Approval of initial Dublin Core Interoperabiity Qualifiers]

From: Larry Masinter <LM@att.com>
Date: Tue, 09 May 2000 00:36:36 +0900
Message-Id: <4.2.0.58.J.20000509003627.0308b440@sh.w3.mag.keio.ac.jp>
To: uri@w3.org
Since Ray is obviously concerned, I'd suggest we invite him to
participate.

 > I'm not sure what you mean by "disarray" but the vast majority
 > of the W3C web site is "personal opinion without critical review."

I just mean that http://www.w3.org/Addressing/schemes.html is
messy, is missing important information, is hard to read, and that
if http://www.w3.org/Addressing/ were clearer, then maybe Ray wouldn't
say "the meaning of "URI" is clearly the subject of pervasive confusion."

 > > There's no longer a link to http://www.w3.org/Addressing/
 > > from  http://www.w3.org/Architecture/
 >
 > Yes, there is; the link text is
 > 	Naming and Addressing (URIs)

Sorry, you're right. It came after all of the listings of 5 years
of history of events in the architectural domain, under the heading
of "Research Notebook", so I missed it. I kind of expected something
about URIs to be part of the Architecture of the web, and not just
a Research topic.

 >> any more, and the addressing
 > > document's not in good shape. For example,
 > >
 > >   A Beginner's Guide to URLs
 > >         The classic intro to URLs, by The NCSA Mosaic team
 > >
 > > is hardly a useful guide to URLs at this point in time;
 >
 > No? Why not? By google's unbiased metrics
 > (cf http://www.google.com/why_use.html), it's the
 > 2nd most relevant page in the Web to the query "URL introduction"

Please don't tell me that you have fallen into the common practice
of confusing "popular" with "good".

 > > and
 > > the document includes the continued confusion over the fact that
 > > syntactically, URNs fit within the URL syntactic space.
 >
 > It is a fact that each URN, for example,
 > 	urn:abc:def
 > fits within the URL syntactic space. I don't understand what you
 > mean by "continued confusion". Please explain in detail.

I mean that the diagram shows that there might be part of the URN
namespace that doesn't fit within the URL syntactic space.


 > > http://www.w3.org/Provider/Style/URI seems to be based on
 > > willful ignorance of economic reality-- that, for example,
 > > companies don't change their domain names when they change
 > > their company names, or that document authors might have
 > > some control over the domain names available.
 >
 > We're completely aware that companies change their domain names;
 > but we maintain that it's cost-effective to maintain the old addresses
 > rather than to break the incoming links.

Yes, of course. The problem is the leap from encouraging a behavior
to designing systems that only work if the behavior is universal.
It would be great if URLs didn't change, and of course you should
encourage it. It's unfortunate, though, that using URLs for namespace
identifiers and other permanent architectural elements _requires_
URLs that don't change.

 > There are likely to be reasonable exceptions, but the Web
 > could hardly suffer from a few more webmasters keeping
 > this guideline rather than breaking incoming links
 > for no good reason.

I don't mind the guidelines at all, they're great. It's just not
robust to have the architecture somehow rely on them always being
followed.

 > > http://www.w3.org/Addressing/schemes.html looks like some
 > > notes that someone (Dan Connolly?) keeps in his spare time.
 >
 > Again, like most of the W3C web site, it's maintained
 > in an ad-hoc fashion, yes. But it's useful to me and
 > at least a few other people, and I have never seen
 > anything elsewhere in the Web that gives this
 > information, so I keep it around.

I'm not at all trying to minimize the utility of this list! I think
it's great! I just think that we should have a little more effort
in getting it into shape, being careful about the status of things,
getting some clarity on the elements that are marked with "???", and
distributing the editorial authority over the contents.

 > > Probably if we got together we could put it into shape
 > > in 3-4 hours of joint work.
 >
 > There's a standing invitation:
 >
 > 	"If you know of any that aren't here, please tell me via
 > the relevant forums."
 > 	-- http://www.w3.org/Addressing/schemes.html
 >
 > By all means, if you have contributions to the page,
 > send them here (to uri@w3.org).

I don't just mean adding new URL schemes, I mean being more systematic
about annotating the schemes that are there, being clearer about their
status, implementation, compliance with various standards and guidelines,
or utility.  I'm willing to spend some effort on this, but I'd like
some other folks to join in, and make sure that there was general agreement
(or at least to document those things for which there's some disagreement.)

 > > I propose that we institute a W3C Activity that consists of
 > > a URI coordination group that meets once to clean up the
 > > existing material, and then meets quarterly (every 3 months)
 > > for a maximum 1-hour phone conference to reviews the status
 > > of the W3C web pages on URIs. Any standards work, documents,
 > > etc, should be processed as IETF RFCs. In addition, we need
 > > someone on W3C staff who has write access to the web pages and
 > > is willing to spend ~30 minutes every 2 weeks updating the
 > > web pages based on comments. I'd establish the write policy
 > > at the initial meeting, but my proposal is for inclusion rather
 > > than exclusion.
 >
 > Well... given that we do this maintenance already, more or
 > less, I'm interested in the possibility of more involvement from
 > the W3C membership; I haven't managed to make a good case for it yet.
 > A quarterly teleconference is an interesting idea, but I don't
 > see obvious criteria for membership; do you have something in
 > mind?

I was thinking we would follow the normal procedure for W3C working
groups, complete with member involvement and invited experts. The
only distinction is that this would be a relatively low level activity.

Larry
Received on Monday, 8 May 2000 14:04:53 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 13 January 2011 12:15:27 GMT