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

Re: Media types

From: Tim Berners-Lee <timbl@w3.org>
Date: Mon, 14 Jan 2002 13:09:53 -0500
Message-ID: <019c01c19d26$aabb8a60$84001d12@w3.org>
To: <www-tag@w3.org>
The core question here is:

> Q. 3; What is, or what should be, the relationship between a media type
> and an XML namespace?

I believe personally that

a)   The sort of document you have is for an XML document defined
fundamentally
 by the namesapce of the outermost element.  Examples:
- The meaning of a document whose outermost element is in XHTML is the
meaning which would be exepcted to be understood by a human user presented
with it according to the spec.
- The meaning of a document whose outermost element is RDF is the semantics
of the RDF statements defined by the RDF spec and eth specs of the
properties used in the outermost level of the RDF document.
The significance of any nesting of one withing the other is to be defined by
the nesting (outermost) specification, although of course some consitency
about this would be great!

b) It is essential for a something handling a message (in Roy Fielding's
terms, a Representation) to know it the body is encoded in XML, as it must
be possible to build systems which parse the XML and then act as a function
of the namespaces.

c) It is useful for anyone handling a mail message or HTTP request to know
the namespace
 (and/or element name) of the outermost element of an XML document.  This
transprency
  just gives more power, reuse, etc.  Examples:
  -  OS desktop uses appropriate Icon
  - Some namespaces will have completey separate applications to launch in
same (currently many) cases

d) For a document with multiple namespaces, it is in general not possible
for a generic program to determine what functionality is needed to handle
it.  It is not possible to determine in which way the namespaces within are
used.  because everything is controlled directly or indirectly from the top
element, one has to instantiate a handler for that and ask it, possibly
recursively.

e) Namespaces, because they are identified by URIs, can have very different
properties of persistence, global agreement, etc.   They can vary between
widely adopted standards and local tranient automatically generated things.
Therfore, one cannot demad that every namespace have a corresponding media
type.

Considering which things,  I suggets that for a namespace wich will be a
widely adopted standard and will be used as an outermost element of a
document, it is wise but not essential to make a special MIME type. If this
is done, though, I would say it is essential that the "+xml" conventions
from the RFC____ be used so that a generic XML processor can be invoked.

I also think it should be emphasised that when a document is simply labelled
as text/xml but uses, at the outermost element, a well-known standard such
as XHTML, SVG, SMIL, etc, that any application which purports to support
that standard hande the file appropriately, and hopefully in an identical
way to the the way it would handle it had a more specific MIME type been
used.

My 2c.

Tim Berners-Lee
IMHO
Received on Monday, 14 January 2002 13:09:55 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:03 GMT