At 9:00 AM 10/29/93 -0800, Ned Freed wrote:
The message from Steve which you characterized as a entirely accurate precis of
remaining issues stated that inclusion of such a parameter made header-set
worthless.
That wasn't what I meant to say. I meant to say that there is very little
difference in my mind between:
multipart/foo
and
multipart/header-set; type=foo
*ONLY* as far as reducing subtype proliferation. From a "how many subtypes
do we have in the world" perspective, "foo" is in either case effectively a
multipart subtype. (Whether or not subtype proliferation is bad is of
course another question.)
But I think you still have the advantage of some sort of hint regarding the
first part of the multipart type (though I'm going to concede that this is
only true of some sorts of header-sets), and the advantage of not having to
repeat the data part just to get alternate representations of the header
part.
--
Steve Dorner, Qualcomm Inc.
Sometimes it's easier to let your cockatoo eat your shoes
than to hear him scream.