W3C home > Mailing lists > Public > public-xml-core-wg@w3.org > January 2010

Re: Minutes for XML Core WG telcon of 2010 January 27

From: John Cowan <cowan@ccil.org>
Date: Thu, 28 Jan 2010 10:29:13 -0500
To: Jirka Kosek <jirka@kosek.cz>
Cc: "Grosso, Paul" <pgrosso@ptc.com>, public-xml-core-wg@w3.org
Message-ID: <20100128152913.GB27323@mercury.ccil.org>
Jirka Kosek scripsit:

> We are in the web space where resources are identified by media types.
> As majority of schema languages are represented in XML and don't have
> specific media type, they use application/xml. But this is fine,
> xml-processor can fetch this resource and from the namespace of root
> element determine whether schema is W3C XML Schema, RELAX NG schema,
> Schematron schema or whatever. This mechanism would fail only for
> non-XML syntaxes like RELAX NG compact syntax, but those formats have
> their own media type.

While this is perfectly true, it is undesirable for a processor to
fetch over the network a schema in a format it does not understand.
This takes additional time for the processor, puts additional load on
the network, and (in the typical case where a large number of documents
refer to a single schema) puts great load on the server which is providing
the schema.

The case of stylesheets is not really parallel.  All major and many
minor browsers support both CSS and XSLT, and no other stylesheet format
is in wide use.  Having a reliable format hint is therefore not really
necessary.  But in addition to W3C XML Schema, RELAX NG, and Schematron,
there are the smaller languages of ISO DSDL to consider -- in short, the
schema language universe is much more fragmented and likely to remain so.
The ability for document authors to provide a robust hint about the nature
of the schema they are pointing to is therefore much more important.

> One can argue that it is quite strange to request access to resource in
> order to determine its content. But in reality most documents will use
> just one xml-model processing instruction so the cost of such check will
> be small. 

I disagree, for the reasons given above.

> As to adding new dedicated attribute -- what was ideas about its
> content? Namespace of the schema language used?

That seems plausible to me.

John Cowan <cowan@ccil.org>             http://www.ccil.org/~cowan
It's like if you meet an really old, really rich guy covered in liver
spots and breathing with an oxygen tank, and you say, "I want to be
rich, too, so I'm going to start walking with a cane and I'm going to
act crotchety and I'm going to get liver disease. --Wil Shipley
Received on Thursday, 28 January 2010 15:29:44 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:16:41 UTC