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

Re: Namespaces and DTDs

• From: Charles Reitzel <creitzel@m...>
• To: xml-dev@i...
• Date: Thu, 11 Mar 1999 11:54:32 -0500 (EST)

Ah, my favorite thing to hate about XML <grin>.

Seriously, though.  I have yet to hear of a single real application that
needs element level prefix declarations.  Not one!  The PI was just fine for
99.99% of applications.  The 0.01% should simply not use XML (or may need an
additional layer, such as AF or a schema processor).  Element declared
namespaces is a solution in search of a problem.  Unfortunately, namespaces
have effectively killed DTD validation.

My wish list for namespaces is as follows:
1) The prefix should be set by document author, *not* the DTD author.
2) The FPI should be set by the DTD author.
3) Prefixes should have document scope.
4) Namespaces should be part of XML proper and *not* an add on.
5) Element names should be resolved in the namespace of the nearest ancestor. 

Until most of these conditions are met, I predict the demise of DTD's.  It
may be too late already...

Best regards,
Charles Reitzel

>From: james anderson <James.Anderson@m...>
>Date: Wed, 10 Mar 1999 11:49:51 +0100
>Subject: Re: Namespaces and DTDs
>
>That "REC-xml-names-19990114" does not provide any means to establish
>prefix<->uri bindings for a DTD has long been a point of contention. A cursory
>search of the archives will bear this out. The decision to eliminate the
>combined prefix/uri/dtd binding (the original pi form) was, however, correct,
>as the pi form, at least as proposed in "WD-xml-names-19980327", would not
>have been sufficient to handle such things as a dtd which needs multiple
>prefix bindings or the situation where a given prefix<->uri binding is to
>apply to multiple schema sources.
>
>While it is true that some mechanism is necessary, a form - as discussed below
>- - which effected a singular binding would also not have solved the problem.
>"Everyone" would seem to be waiting for "schemas"....
>
>Marc.McDonald@D... wrote:
>> 
>> A simple extension to namespaces could have fixed this problem:
>> 1.      Allow a DTD to be optionally specified along with the namespace
>> prefix and URI
>> 2.      When an element is prefixed, parse it using the DTD associated with
>> the namespace and the given prefix as the default.
>> 3.      If no DTD is associated with the prefix or not validating, do what
>> is done now (ensure element is well-formed).


xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev@i...
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ and on CD-ROM/ISBN 981-02-3594-1
To (un)subscribe, mailto:majordomo@i... the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo@i... the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa@i...)

• Follow-Ups:
  • Re: Namespaces and DTDs
    • From: David Megginson <david@m...>

• Prev by Date: Re: ModSAX: Proposed Core Features
• Next by Date: Re: Namespaces and DTDs
• Previous by thread: Re: Namespaces and DTDs
• Next by thread: Re: Namespaces and DTDs
• Index(es):
  • Date
  • Thread