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

 


Help: OASIS Mailing Lists Help | MarkMail Help

 


 

   [] DTD validation [can] work with namespaces []

[ Lists Home | Date Index | Thread Index ]
  • From: james anderson <james.anderson@mecomnet.de>
  • To: "Winchel 'Todd' Vincent, III" <winchel@mindspring.com>
  • Date: Mon, 14 Aug 2000 18:26:08 +0200


----------
nb. this is not a note about the kinds of sematic baggage with which
namespaces might be encumbered. it is also not a note about whether a
namespace name can contribute to identifying the location of a dtd. it
is a note about how to encode multi-dtd documents in a manner which a
not-radically extended xml 1.0 parser could be made to parse.
----------

if one adds the requisite attributes to the combining dtd, it is
possible to infer the namespace membership of the names which one wishes
to capture.

>   <!-- A DocBook/HTML DTD -->
>   <!ENTITY % docbook PUBLIC "-//OASIS//DTD DocBook XML V4.1.1//EN"
>                      "some system URI here">
    <ATTLIST table
       xmlns  CDATA #FIXED "-//OASIS//DTD DocBook XML V4.1.1//EN">
>   %docbook;
> 
>   <!ENTITY % html PUBLIC "-//W3C//DTD XHTML V1.0//EN"
>                      "some system URI here">
    <ATTLIST table
       xmlns CDATA #FIXED "-//W3C//DTD XHTML V1.0//EN">
>   %html;
> 


while this is not a valid xml-1.0 encoding, it is concistent with
section 4 (using qualified names) of the namespace recommendation. it
also provides the parser with sufficient information to intern names in
the proper namespaces iff the parser is using the attribute defaults as
it decodes.

"Winchel 'Todd' Vincent, III" wrote:
> 
> / "Winchel 'Todd' Vincent, III" <winchel@mindspring.com> was heard to say:
> | I suggest to Marcus).  This would allow you to use elements from someone
> | else's DTD without having to create one big DTD and you might also be able
> | to take advantage of some of the Namespace features (defaulted prefixes,
> for
> | instance), which I can't see how you would do otherwise.
> 
> <NormanWalsh>
> I don't think it's a problem to require the author who wants to mix
> several DTDs together to build the compound DTD. It's pretty straight
> forward, and it lets all of the existing machinery (modulo the namespaces
> tweak) to work the way it does now:
> 
>   <!-- A DocBook/HTML DTD -->
>   <!ENTITY % docbook PUBLIC "-//OASIS//DTD DocBook XML V4.1.1//EN"
>                      "some system URI here">
> 
>   %docbook;
> 
>   <!ENTITY % html PUBLIC "-//W3C//DTD XHTML V1.0//EN"
>                      "some system URI here">
>   %html;
> 
> That's pretty short and easy to do.
> </NormanWalsh>
> 
> This is not correct.  The problem --> Element collision.
> 
> Further, as more or more DTDs are developed, there will be more and more
> element collision.
> 
> Classic example: An HTML <Table> and a furniture manufacturer's <Table>.
> 
> Consider the legal industry -- any document, from any source, can show up on
> a lawyers desktop as a peice of evidence.  This means mixed namespaces.
> 
> Thanks,
> 
> Todd





 

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

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