From: Charles Lindsey (chl@clw.cs.man.ac.uk)
Date: Fri Aug 10 2001 - 06:42:27 CDT
In <yl1ymlt18a.fsf@windlord.stanford.edu> Russ Allbery <rra@stanford.edu> writes:
>I should point out here that this problem could potentially be fixed by
>standardizing group aliasing and a return code to POST that said that the
>post was rejected because it was posted to an aliased group, allowing the
>posting agent to itself modify the Newsgroups header appropriately and
>retry, potentially configurably without human intervention by the poster.
The current NNTP draft recognises 'y', 'm' and 'n' in the LIST ACTIVE
command; no mention of 'x' or '=newgroup'. It that were rectified, then
there could indeed be a return code to POST that says "refused because the
active file says 'xxxx'" (where 'xxxx' would include the name of the
newgroup - or indeed any other nonstandard string).
>May or may not be worth the bother.
It could be raised on the NNTP list if people here think it would be
useful. It would certainly provide some future flexibility.
>> Anyway, I posted some concrete proposals a couple of days ago. Nobody
>> has responded yet.
>I'm still back at "I'd rather not include experimental protocols in a
>standards-track document," and beyond that haven't had much time to think
>about long-term future-of-news sorts of stuff lately.
Yes, but first you have to decide whether what I described amounts to an
"experimental protocol", since it was little more that 'mvgroup = newgroup
+ rmgroup', with some MAYs to encourage future developments.
My problem with an "experimental protocol" in this case is that if our
standard does not bless it in even such a rudimentary form, nobody is
going to take any notice of the eventual 'mvgroup' RFD because they will
say "why should we (hierarchy admins) bother using this new header when we
know of no sites supporting it"; at the same time, sites will not be
implementing it because "no hierarchy admins are using it". You need some
pressure from our draft to break that deadlock. After that, it will
prosper, or not, on its merits, and if it starts to get used, then there
will be pressure to implement the MAY parts of it properly.
In this respect 'mvgroup' is different from 'Replaces', which might indeed
start out with small-scale usage, with implementations appearing as the
demand grew.
-- Charles H. Lindsey ---------At Home, doing my own thing------------------------ Tel: +44 161 436 6131 Fax: +44 161 436 6133 Web: http://www.cs.man.ac.uk/~chl Email: chl@clw.cs.man.ac.uk Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K. PGP: 2C15F1A9 Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5