XML and Objects

[Rick Jellife]

david at megginson.com ¼g¹D¡G

>  This sounds like a wise choice.  XML packaging is a problem that the
> W3C XML Activity has not yet addressed -- experimentation and
> implementation experience will be very helpful to them when the time
> comes (after all, that's one of XML-DEV's greatest strengths).

The XML WG made the definite choice to avoid specifying a packaging solution for
MIME, when Murata and Whitehead put together the RFC for text/xml and
application/xml.

I think this was because:

* they wanted to provide the base-level MIME types first and fast: HTML has gone
far with just text/html;

* the URLs locate resources (i.e. entities) not documents per se: so just a MIME
type for entities is not inappropriate;

* there was considerable fuss with text/sgml (one side says that the document
should be parsed first, and only the entities that are actually referenced should
be sent, in the order they are required; the other side says the server should be
able to bundle anything it wants into the package, and that no transitive closure
is required) and the WG needed to sidestep it;

* multipart XML documents can be sent using text/sgml (oops I am doing this from
distant memory...) anyway, if you send the appropriate SGML declaration with it;
so there is already something available, even if it is not optimal;  text/sgml
lets you ship SOCATs too I think;

* in any case, packaging is more appropriate for email than browser delivery, and
XML was "SGML on the Web" not "SGML over Email", so perhaps there is no strong
requirement for the WG to provide a solution, even if there is a gap;

* because SGML documents may be shipped with just public identifiers on entity
declarations, SGML documents may require a SOCAT file, so text/sgml needed to
look at the multipart issue--XML documents must have system identifiers on entity
declarations, hence they need not require a SOCAT file, hence the multipart issue
is peripheral to basic text/xml on MIME on HTTP.  If an XML document does use and
require a SOCAT, then the developer of the document system has to figure out how
to arrange it to work with MIME HTTP.

There has been some ISO discussion about this issue. I would be very interested
if anyone on XML-DEV has any fresh perspective on this.

Rick Jelliffe


xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)