W3C home > Mailing lists > Public > www-zig@w3.org > March 2001

Re: Next ZIG - new record type

From: Mike Taylor <mike@tecc.co.uk>
Date: Fri, 9 Mar 2001 10:32:02 +0000
To: rden@loc.gov
CC: www-zig@w3.org
Message-Id: <E14bKBm-0001pR-00@fw-smtp.tecc.co.uk>
> Date: Thu, 08 Mar 2001 18:06:17 -0500
> From: Ray Denenberg <rden@loc.gov>
> A namespace identifier is not intended to retrieve the namespace, it
> is intended to identify a namespace definition. (If you receive a
> namespace identifier and you don't have the definition, then you
> need to resolve it - but in a very crude sense that's analogous to
> the way oids work.)

Right.  (For those who didn't spot it, Ray is now talking about the
rather idiosyncratic way that XML Namespaces use URLs as unique
identifiers for identifying namespace definitions.  See the W3C
document at http://www.w3.org/TR/1999/REC-xml-names-19990114/#ns-decl
or (perhaps more usefully) "Namespace Myths Exploded" #1 at

> And of course we're talking about what UIRs *should* do, not what
> (http) uris do now.  Hopefully sometime in the next few years
> they'll get this sorted out and URIs will really do these
> things. But in the meantime, I don't have much enthusiasm for
> junking oids in favor of http URIs.

No-one's proposing junking OIDs (are they?) -- just allowing URIs to
be used alongside them when appropriate and convenient.

Much as I tend to be a reactionary old git every time someone proposes
radical upheavals in Z39.50, I really don't think that there is too
much wrong with this one.  It's certainly true that URIs -- even
boring old HTTP URLs -- can be used today as OID-like abstract
identifiers, and their use to identify XML Namespaces proves this by
demonstration.  I also think that Ray's URI:OID:1.2.840.10003.13.8
arrangement makes a lot of sense.

The real issues here are surely to do (once more) with backwards
compatibility.  Allowing InternationalString URIs everywhere we
currently allow OIDs would be a big deal.  As always, Sebastian talks
a great deal of sense when he proposes that we fault in the URIs where
we find real needs for them -- which so far is only in the schema
specification for retrieval.

 _/|_	 _______________________________________________________________
/o ) \/  Mike Taylor -- <mirk@mail.org> -- http://www.miketaylor.org.uk/
)_v__/\  "It doesn't matter if you win or lose ...  Until you lose" --
	 Angie Papadakis.
Received on Friday, 9 March 2001 05:34:29 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:26:02 UTC