W3C home > Mailing lists > Public > www-xml-schema-comments@w3.org > January to March 2001

Re: CR-56: Specify text/xml or application/xml

From: <muraw3c@attglobal.net>
Date: Thu, 25 Jan 2001 01:35:24 +0900 (JST)
To: ht@cogsci.ed.ac.uk
Cc: marting@develop.com, www-xml-schema-comments@w3.org, connolly@w3.org, ricko@gate.sinica.edu.tw, muraw3c@attglobal.net
Message-Id: <20010125013524V.muraw3c@attglobal.net>
Henry,

> To clarify a bit, our decision was based on widely shared experience
> of difficulties with server configuration.  It is one thing to say
> that XML Schemas are in some intrinsic sense application/xml or
> text/xml (sorry, as you know I disagree with the basis for the whole
> text/... application/..., but that's a _personal_ disagreement and
> _not_ the basis of the WG's decision), it is another thing to require
> conforming processors to refuse to process something at the end of a
> URL for a schema document because the server at that end delivers it
> as text/plain.  We did not feel it would serve any good purpose to
> reject the document under such circumstances.

I do not think that RFC 3023 or XML 1.0e2 make the XML processor to
reject XML documents which are not labelled as text/xml,
application/xml, or */*+xml.

Neither RFC 3023 nor the XML recommendation says anything about the
behavior of the XML processor when the media type is not an XML media
type.  The XML Schema spec can simply follow this tradition.

On the other hand, "application/xml" should be mentioned as an 
appropriate media type.
 
> Would it make the decision we took acceptable if it were made clear in
> the spec. that XML Schema documents _should_ be served as ???/xml, or

They should be served as application/xml.  (If this sentence is added, 
I will withdraw my request to the W3C director.

> even that XML Schema conformant processors should be capable of being
> invoked in 'strict MIME' mode in which they would reject documents
> which are _not_ ???/xml?

I do not think that this is necessary.

> How is your request consistent with expected advent of
> application/xml-schema?  Do you expect that we will rev the XML Schema 
> REC (as we hope it will be by then) when this happens?

I am not really sure if application/xml+schema is required.  (BTW, the
delimiter is now "+".)  But there are some possible reasons.  First,
you might want to introduce fragment identifiers specific to XML
Schema.  Second, you might want to do negotiation (e.g., give me
either DTDs, XML Schema, RELAX, TREX, Schematron, etc.).  Third, you
might also want the WWW browser to invoke a viewer for XML Schema
without parsing the document.

When such a media type is created, the XML Schema spec should 
mention that media type as well.

Hope this helps.

Cheers,

IBM Tokyo Research Lab &
International University of Japan, Research Institute

MURATA Makoto
Received on Wednesday, 24 January 2001 11:36:13 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Sunday, 6 December 2009 18:12:49 GMT