At 04:22 PM 3/17/00 -0800, Paul Hoffman / IMC wrote:
For the record, I don't know why the authors of IOTP chose to use a
different sub-tag; they may have a very good reason. But my guess is that
most XML-based applications that want to be found by generic XML parsers
SHOULD use text/xml and application/xml.
I repeat: whether the inventors of the media type, today, design it
for dispatch to generic XML processing machinery (and most won't) is
irrelevant to the point here. The core notion is that the knowledge
that some media type happens to use XML syntax opens the door to
unanticipated kinds of processing beyond those envisioned by its
inventors, and on this basis identifying such encodings is a good
and useful thing.
Simon St. Laurent wrote
I think we have another fundamental disagreement here. Such 'specialized
systems' are already distributed by the millions, in the forms of IE 5 and
Mozilla. Right now, they have pretty limited capabilities, but they're a
start.
Yes, we disagree here. There is no reason to automatically dispatch
application/foo-xml in these XML systems.
Up till a year ago, I might have agreed. As of now, there is a lot of
stuff floating around the net that claims to be XML. For a lot of it,
I don't have the appropriate receiving software on hand; in these cases,
I've found it tremendously useful to throw it into IE5, which first of
all, tells me quickly whether it really is XML or not, and secondly, allows
me to look at it in not-too-unreadable form. I would *love* it if I
could configure my user-agent to, whenever it sees a media type it
doesn't grok but happens to look like foo/bar-xml, to turn it over to the
generic XML-capable browser. [Or even better, based on some other mime-level
mechanism more elegant than the "-xml" suffix.]
Tim Bray gave a few far-fetched
but possibly valid ones, but all of them would need to be used content
sniffing even if we had -xml because each one really, really wanted to be
sure to get all XML.
Far-fetched is a matter of opinion; but the conclusion doesn't follow; one
of the virtues of XML is that if data which purports to be isn't, you
find out quickly and with deterministic and easy-to-handle results.
This doesn't mean that it's a good idea to *always* check opaque media
types for XML-ness; XML processors are not as lightweight as they should
be (sigh, mea culpa) and I don't think you want to poking around at
the beginning of everything that's going through. -Tim