[Zope] - Re: XML-RPC

[Christopher G. Petrilli]

On Sat, Jan 23, 1999 at 01:25:07PM -0500, skip@calendar.com wrote:
> [ I point out the difficulty in negotiationg text/xml with
>   applications ]
> In reverse order:
> 
> I suppose text/xml is either already approved or well on its way to
> approval.  It's too hot for the approval folks to ignore.

Alas, yes you are right, bu then IETF has been more driven by the suits
than engineering since IPv6 started... What won won because (in my
opinion as one of the authors of a competing solution) cisco/etal didn't
want anything that might require HUGE rewrites... even though it would
scale much further, longer, and allow for much faster routing...

> [ Skip compares text/xml to image/gif, and notes no application layer
>   symantics are present ].

This is a bit of a red-herring... there is a substantial difference that
comes from the following:

XML really requires (at least SGML) a DTD to form it, and this is
carried in the body, but in many things (according to my friends at
Interleaf) it's implicit and the DTDs will go away over time...

Now, how am I to do negotiation then to say that I accept XML for a
specific purpose?  Perhaps I accept an XML doodad for purpose A, but not
purpose B, but if I say I accept MIME type 'text/xml' then I can't
differntiate between these. 

Your comparison to GIFs isn't material, as it's content, not format
driven.  There are different MIME types for Real Audio v. QuickTime,
even though you could describe both as a single type of octet-stream,
no?  XML can vary all over the place, and XML-RPC has a specific
specification, therefore, it should have it's own bit attached.

text/blah to ME is reserved for that information destined to be human
readable, which most certainly XML is not.  Therefore, the comment about
xml/foo, not text/xml.  But that's just me, I have no participation in
IETF any more.

Chris
-- 
| Christopher Petrilli
| petrilli@amber.org