On Fri, May 18, 2001 at 08:33:50AM +0100, Gervase Markham wrote:
> The Accept: header pref support is now checked in and working. This makes
> it much easier for us to change the Accept: header in future following
> this consultation process. The current header is:
>
> Accept: text/xml;q=1, text/html;q=0.9, image/png;q=1, image/jpeg;q=1,
> image/gif;q=0.9, text/plain;q=0.8, text/css;q=1, */*;q=0.01
cool. I don't have a good answer for you at this point but it has been
raised in the correct forum. In the mean time, you should study CC/PP
which is a general solution to negotiating client/server/proxy
capabilities. See Philipp Hoschka's mail of Sun, 13 May 2001 10:26:53
+0200 Message-ID: <[EMAIL PROTECTED]> for more description and
pointers.
> Brendan has made the point that MNG needs to be in there; we also need
> application/xml and application/xml+xhtml too.
I think you mean application/xhtml+xml and probably text/xhtml+xml.
RFC 3023 [1] discusses the +xml use and semantics. I don't believe
html+xml or xhtml+xml subtrees have been defined. SVG is an example
of the +xml hack as it has a content type of "Image/svg+xml". This
is a compelling example as the "+xml" suffix allows the interpreting
application to handle the data as XML, changing the tree from image
to text in the process. You may find "XML Document Interpretation" [2]
usefull in further study.
> Eric, Philipp - who exactly do we have to ask or where to we have to post
> to get W3C input on a suggested header? We need to bear in mind the
> requirement to keep it short.
The IETF is responsible for "registering" MIME headers (like
Content-length) and HTTP headers (like Connection or Server). I
beleive the authoritative list is the RFCs; all you need is grep (for
no particular string). You may want to investigate HTTP Extensions [3]
which provides a mechanism to uniquely identify headers that have not
been "registered" in this fashion.
> One other question: is application/xml supposed to be treated as a more
> general (as in the relationship between image/* and image/gif) form of
> types like application/xml+xhtml ? How does that work?
I think RFC 3023 provides the answer, which is that */*+xml, text/xml,
application/xml (like text/xml except that the content provider
requests that the data not be presented to the user as text),
text/xml-external-parsed-entity (DTD subset), and
application/xml-external-parsed-entity (DTD subset) all represent XML
MIME entities. From RFC 3023 section 7:
[*/*+xml] will allow applications that can process XML generically
to detect that the MIME entity is supposed to be an XML document,
verify this assumption by invoking some XML processor, and then
process the XML document accordingly.
[1] http://www.imc.org/rfc3023
[2] http://www.w3.org/2001/03/06-XML-document-interpretation
[3] http://www.w3.org/Protocols/HTTP/ietf-http-ext/
--
-eric
([EMAIL PROTECTED])
feel free to forward this message to any forum.
