At 12:14 PM 3/17/00 -0500, Simon St.Laurent wrote:
I'm wondering how this kind of content negotiation might or might not
integrate with the use of the -xml suffix, and how it might or might not
work if that suffix isn't used.
In my opinion, trying to base content negotiation on this kind of suffix
takes us over the threshold from a "mostly harmless" but often useful way
to recognize XML for generic handling into a realm of potential complexity
and problems of the kind that Pete and Keith have been concerned about.
IMO, the current proposal (-xml suffix) will never be 100% perfect: there
may be content types containing XML that don't use the -xml extension. For
the purposes of recognition for generic XML processing this doesn't seem
like too much of a problem. And environments that don't have any concept
of generic XML processing can simply ignore the naming convention.
To incorporate such a mechanism to protocol developments that really should
try to achieve 100% accuracy, and to enshrine the naming conventions into
future protocol developments is not, in my view, the way to go.
Once we start looking to content negotiation protocols, I'd suggest looking
at indicating generic XML-ness through a CONNEG style media feature expression.
#g
------------
Graham Klyne
(GK(_at_)ACM(_dot_)ORG)