- From: Dan Connolly <connolly@w3.org>
- Date: Tue, 11 Feb 1997 17:57:16 -0400
- To: David Durand <dgd@cs.bu.edu>
- CC: w3c-dist-auth@w3.org
David Durand wrote:
>
> The W3C position, strongly
> articulated by Dan Connolly, was that any interpretation of URLs was not
> acceptable.
Just to be clear: That's certainly my position (and TimBLs and
Henriks)...
calling the W3C position might be
overstating things a bit -- we aren't speaking for the members
in this case.
> As above, "URL-hacking" was a controversial approach that some were not
> willing to accept. The functionality is crucial, but if URLs cannot be
> interpreted new HTTP methods are needed, and a round-trip is required for
> certain obvious operations.
Huh? I don't think that conclusion follows: What's wrong with sending
the version info along side the URL whenever you need it, in order
to prevent round-trips? That's what protocol design is all about, right?
<resource href="XXXXXXdon'tlookinsidemeXXXXXX">
<link rel=version-successor href="YYYYdon't/look/in/here/either">
</resource>
That completely tells the relationship between XXX and YYY without
URL-munging.
If you're not familiar with the <resource> concept, please see:
http://www.w3.org/pub/WWW/MarkUp/Resource/Specification
Actually, more folks seem to like the name <about> for this purpose. And
other folks want to express the same thing in S-expressions. Stay tuned
to
see how it all comes out.
The bottom line is: don't look in the name for this info. Put it
somewhere else.
> I will add that VTML also contains facilities to send multiple versions of
> a document in a single reasonably organized package. This is essentially a
> simplification/generalization for general stream of the kind of information
> kept in a version repository.
Yup... that's what this group is doing: figuring out how to exchange
data between version repositories.
Dan
Received on Tuesday, 11 February 1997 16:57:17 UTC