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


Help: OASIS Mailing Lists Help | MarkMail Help



   Re: [xml-dev] more QName madness

[ Lists Home | Date Index | Thread Index ]

John Cowan wrote:

> I do not understand the URIs Good, QNames Bad point of view, since they are
> plainly isomorphic.  Indeed, RDF defines
> a two-way mapping between QNames and (some) URIs.  It would be more consistent
> for you to attack URIs, QNames, and IP addresses, all of which are universal
> agreements creating global names, in favor of some UUCP-like scheme where the
> node that calls itself "deutsch" is known to its various neighbors as "german",
> "duits", "tedesco", "allemagne", etc. etc.

[with apologies in advance to those who hear this from me all too often]

I believe that you are confusing the context of the document (within an
internetwork addressing scheme) with its content, which we may assume exhibits
both integrity and autonomy. [In my own 'best practices' I rigidly enforce the
distinction by using attributes for the first and elements for the second, but
that is a subject for another discussion.] Having a context for a document is the
necessary consequence of having access to that document via the internetwork. By
contrast, we should assume that the content of a document is entirely independent
of its internetwork context. Documents are as their creators create them. A
document changed as its transient context is altered is profoundly not the same
document. REST is built largely upon that single premise.

QNames in content are the mutilation of document integrity to embed context, and
worse, context from a single, particular, transient point of view. If we grant the
autonomy of a processing node, then the context which we have gone through such
contortions to embed as QNames is useless to the very processing nodes where we
had hoped that it would facilitate interop, because that context is unlikely to be
the context of interest from the unique point of view of such a processing node.
Processing on the internetwork inevitably involves fetching, or at least
accepting, documents from internetwork-addressable locations. With each of those
documents comes a context utterly specific to that use of that document on that
occasion. Internetwork interop has to be predicated on the ability of each process
to identify, distinguish and combine those particular contexts into a taxonomy
unique to each instance of process. What could be further from serving this
requirement than demanding an unachievable, static, a priori universal agreement
on namespacing or name-qualifying standards for use in document content?


Walter Perry


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

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