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

Re: Request for Erratum to XML 1.0 and 1.1 Specs

• To: michael.h.kay@n...
• Subject: Re: Request for Erratum to XML 1.0 and 1.1 Specs
• From: Rick Jelliffe <ricko@a...>
• Date: Wed, 22 Oct 2003 20:08:17 +1000
• Cc: xml-dev@l...
• In-reply-to: <005201c39871$c748f6c0$6401a8c0@pcukmka>
• References: <005201c39871$c748f6c0$6401a8c0@pcukmka>
• User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.3.1) Gecko/20030428

Michael Kay wrote:

>I have no objection to anyone trying to find a solution to this problem,
>only to solutions that break architectural interfaces. 
>
After various comments and looking at the specs again, I think that what 
I am suggesting
may be already allowed by the XML specification. So I have modified by 
request to use
different words, but the practical intent is still the same: to allow an 
feature on SAX
parsers so that references to standard entities for characters when 
there is no DTD
will be handled by default, and thus to reduce the dependence on DTDs.

The XML Spec does not say what applications must do with Unexpanded 
Entity Reference
Information Items (UERII) which have no replacement text or system 
identifiers,
from a non-validating XML processor. It explicitly mentions that an 
application is
allowed to replace the reference with the replacement text, so AFAIK we can
assume that an application can do anything it likes.

So the issue becomes
 * the XML Spec should have a little note to make this clear
 * some general cross-application mechanism should be created, and W3C
is the appropriate place for it, and I think the little note in the XML spec
is again the best place for it
 * other standard applications may care to commend this approach
on incoming data (e.g. XSLT, XML Schema, Schematron, RELAX NG)
so that SAX parsers implement it.

The particular approach I think is best is the minimal one, that the 
definitions
from HTML4/ISO and MathML/ISO be used to expand those UERII.

Cheers
Rick Jelliffe

• References:
  • RE: Request for Erratum to XML 1.0 and 1.1 Specs
    • From: "Michael Kay" <michael.h.kay@n...>

• Prev by Date: Re: Alternative "character entity" proposal
• Next by Date: Re: UTF-8+names
• Previous by thread: RE: Request for Erratum to XML 1.0 and 1.1 Specs
• Next by thread: RE: UTF-8+names
• Index(es):
  • Date
  • Thread