Chris Newman (cnewman@iplanet.com)
Wed, 14 Mar 2001 17:02:01 -0800
--On Wednesday, March 14, 2001 11:31 -0500 Eric Burger
<eburger@snowshore.com> wrote:
> OK, here's another choice (thanks to Mark Crispin): How about making the
> critical content notation a parameter to Content-Type?
Poor choice. The scope of Content-Type parameters is intended to be
specific to the content type. Thus a "foo" parameter on application/bar
can have a different meaning from a "foo" parameter on application/baz.
If you want to add a parameter to an existing header, add it to
Content-Disposition which has parameters which are _not_ specific to the
disposition value and is already in the IMAP BODYSTRUCTURE. The "filename"
parameter on Content-Disposition is a perfect example of this sort of thing.
Futher, the concept of "criticality" fits much better as an aspect of the
"disposition" of the body part rather than a modifier to the "type" of a
body part.
If I could go back in history and redo things, I'd replace BODYSTRUCTURE
with a headers/headers-not facility that fetches the skeleton of the MIME
message minus the body parts themselves. Any client which does offline or
local file work has to have an 822/MIME header parser anyway, so why force
clients to implement two syntaxes for the same data?
- Chris
This archive was generated by hypermail 2.0b3 on Thu Mar 15 2001 - 03:05:32 IST