ietf-xml-mime
[Top] [All Lists]

Re: Starting the ietf-xml-mime mailing list

1999-04-08 05:16:33


Larry Masinter wrote:

I tend to lean towards "every application gets a new media type".

In case it isn't obvious by now, so do I.

This is a fine maxim, but it begs the question: what constitutes
a separate, distinct application, vs a modification, profile,
extension, or varation of an old one?

Thats a good question, but part of it comes down to monolighic versus
granular compound documents.

In the case where you're allowed to have a document that mixes
traditional HTML, 

(I'm assuming you mean XHTML here, ie HTML written in XML. Sop, all the
things you citer are written in XML, and constitute one large XML
instance using multiple namespaces).

MathML, Vector Graphics ML, etc., are these
separate "applications" or are they one "application" ("renderable
XML document")?

If sens as separate bodyparts, which is possible, then they would all be
separate "applications". If sent as one big XML instance, then in this
case I would ay it would be text/xml.

Once XLink and XPointer get more widely deployed, we are likely (IMHO)
to see people moving away from this class of monolithic document and
instead using small skeleton documents, written in XML and using
XPointer transclusion to reference different media pieces (graphics,
math, audio, etc) some (many) orf which are writen in XML , but each
with their own label. This granularity allows finer control by mail user
agents (they can figure out which bodyparts they understand) and allows
the benefits of cacheing over HTTP as different compound documents reuse
different component pieces.

So (to get back to Larrys point) while we may see monolithic
multimnamespace documents which can only really be labelled as text/xml,
we will also alongside that see smaller more granular compound
documents, where a correspondingly finer grained labelling of each
component is more appropriate.


Since it doesn't seem to have come up yet, I will also note the concept
of content negotiation, as used by HTTP; where MIME labelling is used in
negotiations between client and server. I guess this could also apply to
email in the case of external bodyparts, but am not sure so will leave
that to someone else. 

Content negotiation can be initiated by the document author (in the case
that multiple alternatives are explicitly offered in the link to the
resource), by the client (in the case of accept headers) or by the HTTP
server (in the case of 'multiple choices' response code). Discussion of
different MIME types for things written in XML should bear these
considerations in mind, as well as the case where the content is already
sent and one is trying to dispatch it appropriately.

--
Chris

Larry