[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

RE: XML Namespace Catalog proposal 2nd edition

• From: Miles Sabin <MSabin@i...>
• To: xml-dev@l...
• Date: Mon, 01 Jan 2001 19:13:57 +0000

Jonathan Borden wrote,
> Regarding the possibility of a single vs. multiple XNC 
> catalogs, the intention is for a SAX entity resolver to [expletive deleted] in 
> an XML Namespace Catalog and resolve off this. One could 
> certainly provide an alternate/local catalog from which a 
> custom entity resolver would resolve. This is an implementation
> issue somewhat orthogonal to the XML Namespace Catalog Format 
> itself.

I'm not so sure.

It's certainly true that a SAX entity resolver would support 
catalog substitution, but that's a pretty blunt instrument. It's 
also fairly opaque: you'd need to see the implementation of the 
resolver to know precisely what the substitution was, and that 
implementation might not be available in any particularly 
digestable form.

By providing references to schemas a catalog supplies at least a 
part of the intended interpretation of a document, and allowing 
for multiple catalogs allows for multiple interpretations.
Allowing for multiple interpretations of a document, and, more 
generally, multiple associations of related resources with a
document strikes me as a Good Thing, and something which should
be supported. But it's not obviously an unqualified good. For
example I can easily imagine XML applications where the authors
of document instances might want to be able to disclaim all
responsibility for the consequences of substituting their
intended semantics (even those semantics which are expressible
via XML schema) with any others.

Even leaving semantics aside, suppose we used catalogs to
associate stylesheets with a document instance (in which case
multiple catalogs => multiple presentations). In some sectors
presentation and changes of presentation are extremely
important, and the authors of a document instance might again
wish to disclaim all responsibility if their intended 
presentation isn't preserved (ie. the authors of a document
representing a legal contract might wish to disclaim 
responsibility if a stylesheet obtained via a substitute catalog 
caused particular clauses to be rendered in extremely small print 
or as white on white text).

So it looks to me as tho' we need something which supports
related-resource substitution with at least the option of that 
being done in a controlled or authorized way. This _might_ be
orthogonal to a catalog format, but I'm not so sure.

I'm open to persuasion tho' ...

Cheers,


Miles

-- 
Miles Sabin                               InterX
Internet Systems Architect                5/6 Glenthorne Mews
+44 (0)20 8817 4030                       London, W6 0LJ, England
msabin@i...                         http://www.interx.com/

• Prev by Date: RE: Namespace related-resource bundles: just the one?
• Next by Date: RE: Namespace related-resource bundles: just the one?
• Previous by thread: Re: XML Display in IE5/Netscape - Urgent
• Next by thread: Namespace Catalogue Hackery (long, mostly code)
• Index(es):
  • Date
  • Thread