W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > January to March 1999

Re: DTD declarations

From: Chris Maden <crism@oreilly.com>
Date: Tue, 23 Feb 1999 18:18:43 -0500 (EST)
Message-Id: <199902232318.SAA20007@ruby.ora.com>
To: w3c-wai-gl@w3.org
[Charles McCathieNevile]
> I have just been looking at the whole thing a bit more carefully -
> it seems to me that you can either use a DOCTYPE or a namespace - if
> you are going to mix bits of different XML then you use
> namespaces. Anyway, I suspect it is an issue we need to sort out.

A DOCTYPE is an assertion of compliance with a data model: in a valid
XML document, all elements' contents match the content models for
their respective element types, required attributes are present, all
attributes are within the types mandated by the DTD, etc.

Namespaces are a way of uniquifying element type names world-wide;
for example, when I say "http://www.w3.org/TR/WD-xsl^template", an XSL
processor knows, "Aha!  This is the 'template' that I know how to
handle, because I recognize that magic string!"  An HTML-Voyager
processor may recognize "http://www.w3.org/TR/REC-HTML5^p" as an HTML
paragraph, but that name makes no assertions about validity of the <p>
nor even that there's actually anything to download at the URI given.

These issues are orthagonal; you can use DOCTYPE declarations without
namespaces, and vice versa; you can use both, and you can use neither.

Now, I don't know which the WAI actually should care about.  On the
one hand, use of namespaces will probably be essential to proper
processing of elements in heterogenous documents: HTML, MathML, RDF,
etc.  And in fact, in such documents, assuring validity is a very hard
problem that hasn't been solved yet; I believe the XML Schema Working
Group is tasked with that problem.  On the other hand, use of a
DOCTYPE does provide some means of testing the structural soundness of
a document, without which it may be impossible to process the document
however well it uses namespaces.

<!ENTITY crism PUBLIC "-//O'Reilly//NONSGML Christopher R. Maden//EN"
"<URL>http://www.oreilly.com/people/staff/crism/ <TEL>+1.617.499.7487
<USMAIL>90 Sherman Street, Cambridge, MA 02140 USA" NDATA SGML.Geek>
Received on Tuesday, 23 February 1999 18:18:46 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 16 January 2018 15:33:29 UTC