From: Russ Allbery (rra@stanford.edu)
Date: Thu Jun 24 2004 - 23:48:01 CDT
Bruce Lilly <blilly@erols.com> writes:
> Russ Allbery wrote:
>> There was a fairly long discussion in news.software.nntp a while back
>> about multipart postings. It's hideously unlikely that anyone will
>> actually use multipart/partial as the primary method of indicating
>> multipart postings on Usenet, at least without *also* relying on the
>> subject hacks, because Content-Type isn't and probably never will be in
>> overview information.
> N.B. that's message/partial, not multipart/partial.
Yup, sorry.
> 1. NNTP != all of Usenet. IMAP is also used, extensively, and doesn't
> have that limitation. [any field or group of fields can be requested]
This is certainly true, although I really doubt that anyone is going to be
downloading any significant quantity of binaries from Usenet via IMAP.
This is one of those things that sounds nice in theory, but it's just not
what people are doing in practice.
> 2. The overview extension to NNTP seems to have a number of limitations.
> However, we can't solve NNTP issues here.
The overview extension to NNTP has a lot of engineering and effort put
into it that appears to be poorly understood by people who haven't
implemented news servers for large sites. :) The path out of this
particular problem for those who really care would be to use HDR, but I
see little to no indication that most of the users of the binary
newsgroups on Usenet have any desire to exert real effort in this
direction, or have enough support on the server side to get much up-take.
> 3. Subject hacks remain hacks. I have seen "[1/23]" etc. used in spam
> on Usenet, where there were no multiple parts associated with the
> message in question. So the practice is highly dubious at best.
> About all we can constructively accomplish is to point out the existing
> standard mechanism for partitioning a large message into pieces and hope
> that the transport folks will take care of relevant matters in their
> areas.
The transport folks are not going to be particularly motivated to put a
lot of work into providing additional protocols that no one will use. :)
-- Russ Allbery (rra@stanford.edu) <http://www.eyrie.org/~eagle/>