OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   RE: Namespaces and URNs

[ Lists Home | Date Index | Thread Index ]
  • From: Charles Frankston <cfranks@microsoft.com>
  • To: 'Peter Murray-Rust' <peter@ursus.demon.co.uk>, xml-dev@ic.ac.uk
  • Date: Tue, 4 Aug 1998 19:25:24 -0700

> -----Original Message-----
> From: Peter Murray-Rust [mailto:peter@ursus.demon.co.uk]
> Sent: Tuesday, August 04, 1998 12:32 PM
> To: xml-dev@ic.ac.uk
> Subject: Namespaces and URNs
> I would appreciate guidance from people knowledgeable in URN 
> syntax and use.
> (A) The NS draft gives an example as
> 	xmlns='urn:loc.govs:books'
> but a private correspondent tells me than URNs should be of the form
> urn:scheme:something - so unless loc.gov is a scheme there 
> appear to be
> differences of opinion.

Note that the examples in the namespace paper are intended to be just that
-- hypothetical examples.  That doesn't mean that the URNs used in the
examples have actually been registered -- they certainly have not.  However,
this example does have a problem -- the period in "loc.govs" is not allowed
per the productions in http://www.isi.edu/in-notes/rfc2141.txt.  The example
ought to at least be syntactically conformant.  I have already pointed this
out, and will ensure it gets fixed before the spec goes to PR.

> (B). I would greatly value consistency in the use of standard 
> URNs. The
> only way we can implement interoperability is if everyone 
> uses the same URN
> to refer to a namespace. Since the URN need not represent a real WWW
> resource there is very little check on whether a URN is fictitious.
> 	The draft shows an example of this problem. HTML is 
> referenced mainly
> through the string:
> 	http://www.w3.org/TR/REC-html40
> This seems fine to me, but later (5, box 6) it is referenced by:
> 	urn:w3-org-ns:HTML
> Are these equivalent? If so, there must be a central registry 
> of synonyms
> somewhere. And, if so, it places a great burden on 
> implementers who must
> search the WWW for all possible equivalences. I would *much* 
> prefer that
> the whole world used exactly the same string for the 
> namespace for HTML4.0.

Again, the examples are meant to be examples, and need not be consistent
among themselves.  As John Cowan pointed out URNs, if and when they are
widely supported, are clearly the best solution for the persistent unique
names of XML namespaces.  Unfortunately, URNs don't really exist yet, and in
interim URLs or other URIs can be used.  

> 	Experience with HTML has not helped the cause of unique 
> identifiers for
> DTDs. I suspect that there are zillions of different (F)PIs 
> in the HTML
> DOCTYPE statements round the world. And - unless I'm told 
> different - I
> suspect no software actually processed them. So we are going 
> into an era
> where we have to convince authors that NS URNs are critically 
> important -
> they can't just guess and hope.
> 	I'd like to suggest that people  who really know post 
> the authoritative
> URNs for common DTDs/schemas. And that we all stick to using 
> the correct
> one in all our examples. Among the ones we shall need in many 
> namespace
> examples are:
> 	HTML		(including version)
> 	XSchema
> 	MathML		(a useful example)
> 	DublinCore
> 	RDF
> 	XML-Data

I think you're confusing URNs with URI schemas.  All URNs use one schema:
"URN".  This list you show above would indicate that you're interested in
defining standard URNs for some popular namespaces.  While this isn't a bad
idea, I think it will ultimately be hopeless to take on the task of tracking
all interesting namespaces.  The namespace spec must allow for the use of
namespaces without requiring each individual namespace be registered prior
to use.  This may mean piggy-backing on existing uniquefying mechanisms such
as Internet Domain names, which would argue for the use of URLs in interim.
This doesn't mean that some of the more popular and widely used namespaces
shouldn't use URNs.  

There is a procedure described in
http://www.ietf.org/internet-drafts/draft-ietf-urn-nid-req-03.txt for
registration of URN namespace ID (the "NID" in the RFC2141 productions)
might be registered.  Unfortunately, the current draft starts out by saying:
'For the purposes of this document, an "IANA-like" entity is assumed to
exist.'  I take this as an indication that the IANA itself has not actually
taken up this task, and that no other organization has stepped forward to do

> Having developed 2 DTDs (CML and VHG) I need to know how to 
> prepare URNs
> for them. It seems clear that anyone wishing to create a DTD 
> which can be
> used for Namespaces has to buy/borrow a domain name. (I 
> suppose they could
> buy an FPI instead, but since I already have domain names I'd 
> prefer to
> stick with those).
> 	P.

Well, you could try to be the first on your block to register a URN, if you
could find who to register it with.

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev@ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo@ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo@ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa@ic.ac.uk)


News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS