W3C home > Mailing lists > Public > www-tag@w3.org > January 2002

Re: Media types

From: Gavin Thomas Nicol <gtn@rbii.com>
Date: Tue, 15 Jan 2002 15:33:41 -0500
To: Paul Prescod <paul@prescod.net>
Cc: www-tag@w3.org
Message-Id: <E16QafS-0005sK-00@server2000.ebizhostingsolutions.com>
On Tuesday 15 January 2002 03:20 pm, Paul Prescod wrote:
> Gavin Thomas Nicol wrote:
> > It really doesn't matter what the specifications say.
>
> If we're trying to answer the question of whether X conforms to type
> Foo then it depends on the definition of Foo. In this case the
> specifications are the definition.

In this statement you are logically testing each and every definition 
of type against the document to determine. My whole point is that you 
shouldn't need to look at the document to determine the processor... 
the document should already have been routed to it by some external 
means.

> So the question is, should the MIME type be
> inferred from the root namespace, some list of namespaces, or
> completely independent?

This is in the right general direction. My  opinion is that 
"type" is purely in the mind of the interpreter... so the question of 
establishing  "type" is irrelevant... we need to determine the 
appropriate processor. In HTTP (and in MIME in general), the MIME type 
is used to determine the *processing* (typically a viewing/playback 
processor, but not necessarily)... so this is one means toward that 
end.

> I don't understand what you are saying. Nobody is talking about a
> universal processor. What I'm hearing is that people want to sniff
> the namespace (either from the MIME header or the XML itself) and
> then *choose* an XML processor based on the namespace. XSLT
> stylesheets with an HTML root-element are one example where this
> would not work.

My point here is that there are an infinite number of ways a given 
document could be processed... so sniffing at elements isn't going to 
help you much with the decision.

I think you need to have some way to associate a means of processing 
with a given peice of data. I don't believe the data can carry this 
information itself (except in a few cases). I also don't think MIME 
type is really sufficient (not granular enough), though some external 
means like it is necessary.

For the moment, MIME type, or whatever other mechanism the transport 
layer provides should be used to infer the processing to be performed. 
In the long run, we need something else.... 
Received on Tuesday, 15 January 2002 16:00:02 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:04 GMT