Excerpts from mail: 13-May-94 Re: Suggest promoting Conte.. Harald
Alvestrand(_at_)uninet (670*)
I can't find Keith's comments where he is "uncomfortable" with the
Difference parameter.
I thought it was mail to the whole list. Unfortunately, I delete mail
ruthlessly, so I no longer have it. I'm sorry if I misattributed the
remark or committed any other faux pas.....
On restricting to Content-* headers:
I don't like adding text to restrict things in more-or-less (to me)
meaningless fashions; if someone comes up with a scheme in which
multipart/alternative; difference=security-classifcation makes sense,
why should I stop them, or try to force the header
"content-security-classification" down their throats?
Well, the main reason is that we've already articulated a general rule
that in parts of multipart messages, gateways in general don't need to
preserve the semantics of headers unless they start with "Content-".
The combination of this with the difference parameter strikes me as
essentially ASKING someone to create a problem wherein we have
"difference=security-classification" but a gateway had nuked the
security-classifcation header from a body part, because it didn't start
with Content-*.
(I feel that the whole business of special treatment of the Content-*
headers is a nuisance and nothing but a nuisance anyway, but that is
another area entirely)
Yes, it is. They're two separate arguments. But I would argue very
strongly that unless we're changing the general rule about Content-*,
we're asking for trouble if we don't use the same restriction for the
difference parameter. -- Nathaniel