W3C home > Mailing lists > Public > uri@w3.org > April 2001

Re: Proposal: 'tag' URIs

From: Sandro Hawke <sandro@w3.org>
Date: Sun, 29 Apr 2001 22:38:08 -0400
Message-Id: <200104300238.WAA01420@hawke.org>
To: "Larry Masinter" <lmnet@attglobal.net>
cc: "Tim Kindberg" <timothy@hpl.hp.com>, uri@w3.org

Larry Masinter wrote:
> Could "dated URLs" meet the same requirements that were identified
> for "tags"?
> 
> http://lists.w3.org/Archives/Public/uri/1998Jan/0009.html
> http://lists.netsol.com/cgi-bin/wa?A2=ind0009&L=URN-IETF&P=R2811

Yes, I think they do meet our requirements as stated, but perhaps we
left out one criterion from that list.

Martin Duerst pointed me to your dated-url proposal, by the way.  Tim
Kindberg and I debated whether to specifically reference it in our
draft, and decided against it, thinking it was more casual a proposal
than one would normally reference and that it addressed a somewhat
different problem.  Instead we simply added a line about there being
various previous proposals in this space.  I apologize if that was
insufficient; I'm still learning the practices in this field.

I like the tag syntax a little more than dated-url, but my main
problem with dated-url for my application is in the semantics of the
URI.  I interpret

    urn:dated-url:200008:http://larry.masinter.net/

as denoting a web page as it existed at some point in the past (even
if there never was such a web page).  (More technically, I would
reference the HTTP RFC's definition of what http: URIs denote.)  Your
e-mail above suggests you mean the string it as a denotation-free
identifier, but I think that's counter-intuitive.  I also think the
web-page-as-it-existed is a very useful concept in its own right.
(Although such a URI scheme probably ought to provide for all the
external variables that could effect the page, not just time.  Complex
issue.)

Maybe I should say a little more about semantics: there's an open
question in the semantic web community about whether http: URIs can
only denote web pages or can be used to denotate arbitrary things in
arbitrary domains of discourse.  One of my goals for the tag: URI is
to provide an alternative, so people who think http: URIs should only
be used for web pages can still use URI syntax for arbitrary
identification.  

   -- sandro
Received on Monday, 30 April 2001 02:41:12 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:25:02 UTC