Lists Home |
Date Index |
- To: firstname.lastname@example.org
- Subject: Re: [xml-dev] Strategies for a lowly XML document
- From: Bill Lindsey <email@example.com>
- Date: Wed, 23 Jan 2002 12:14:34 -0800
- Organization: B-Bop Associates, Inc.
- References: <3C4EDE73.62D3CEA6@mediaone.net>
- User-agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:0.9.4) Gecko/20011128 Netscape6/6.2.1
Jonathan Borden wrote:
> <?rddl-doctype href="http://example.org/some-rddl-description.html" ?>
> I've heard both sides, some people desire "document type" based
> processing, others don't. Is there any harm in providing for this
> _optional_ PI?
My objection is that is another way to accomplish the same goal
that is already part of a W3C recommendation: the "xsi:type"
attribute from W3C XML Schemas.
Your goals, as I understand them:
* Identify a document as belonging to a named type
* Using the name itself as a reference to a resource directory
xsi:type does the first, and as best as I can tell doesn't
preclude our doing the second.
At 1014 23/01/2002 -0500, Elliotte Rusty Harold wrote:
> I don't think it's as clear cut as this. Imagine
> that your application, for
> whatever reason, wants to package several XML
> documents together into one
> overall document. You might then want to write:
> <DocSet rddl:doctype="http://whatever/DocSet">
> <doc1 rddl:doctype="http://whatever/doctype1">
> <doc2 rddl:doctype="http://whatever/doctype2">
What's wrong with:
I'd be particularly interested in hearing
arguments that this approach:
* somehow subverts W3C Schemas or Namespaces
* requires W3C schema validation or PSVI processing
* interferes with W3C schema validation or PSVI processing
* favors any particular schema language over others
* is ugly