[
Lists Home |
Date Index |
Thread Index
]
- From: Ian Graham <igraham@ic-unix.ic.utoronto.ca>
- To: xml-dev@xml.org
- Date: Tue, 24 Oct 2000 18:14:38 -0400 (EDT)
On Tue, 24 Oct 2000, Simon St.Laurent wrote:
> At 01:52 PM 10/24/00 +0800, Rick JELLIFFE wrote:
> ......
>
> Content negotiation could offer an 80/20 solution - but it needs to be
> documented, discussed, and sorted out. The relationship of namespaces to
> resources with content of various different MIME Content Types isn't
> exactly a small thing, and content negotiation in general is pretty weakly
> implemented so far. Organizationally, this is at the boundaries between
> the W3C and the IETF, and philosophically, it brings us back into the
> philosophical murk of 'what is a namespace anyway?'
I would be very much against using content negotiation to do this, as this
just avoids the real problem -- that current negotiation schemes cannot
exchange the information needed to select one of many text/xml variants.
This is in part an HTTP design issue, and I'm sure the HTTP working group
is looking into this sort of issue (teh simplest, and no doubt wrong way
to do this would be to ad some sort of XMLnamespace: header field to the
HTTP header, to define requests/responses for XML 'subtypes'). The
proposal to use text/xml-dtd (etc.) MIME types is an obvious workaround
for this -- but is likely not an appropriate long-term solution.
> Seems to me like it's worth discussion in broad daylight rather than
> implementation by night.
I think you are confusing a simple experiment /attempt to provide a
useful service with standards setting.
Ian
|