From: Charles Lindsey (chl@clw.cs.man.ac.uk)
Date: Tue Aug 14 2001 - 04:30:38 CDT
In <ylzo93o7q5.fsf@windlord.stanford.edu> Russ Allbery <rra@stanford.edu> writes:
>I think you'd need some evidence that people actually use it for useful
>things that would be cool to support, and all the discussion in the world
>is unlikely to provide that evidence. I feel the same way about a lot of
>the things people propose. That's why we need to get the base draft out
>so that we have a working extension mechanism,
Actually, we have not defined an extension mechanism. OTOH, I don't think
we actually need one in the way that NNTP does.
OTOH, although we may have been talking about "experimental" extensions,
please be aware that, it they are envisaged as candidates for
standardisation later on, then they have to be published as "standards
track" RFCs because, according to the rules, you cannot move an
"experimental protocol" RFC onto the standards track without first
republishing it as a "proposed standard" RFC (even if it has not changed).
>because then stuff like
>this can be advertised as an extension and used or not used depending on
>whether it's available, and then people can actually start playing with it
>and we'll have hard evidence on how useful it is.
Yes, but 'mvgrpoup' is not the sort of thing you can "use or not depending
on whether it's available". Either you know it's widely supported (or you
see it in a standards-track document and believe it soon will be so), in
which case you (as a hierarchy admin) use it. Or else you believe only a
minority of sites support it, so you don't bother (and that give no
encouragement to the rest of the sites to implement it).
In any case, suppose we were to issue such an extension a few months after
our present draft is published; would we write into that extension
anything different from what we would write now? Will we understand the
issue any better by then? Or are you saying any such extension must wait
until you are ready to tidy up the 'ugly' parts of INN that are currently
making it difficult?
Extensions are a fine way to proceed in the case of a feature that will be
useful even if it is only implemented on a few sites. Then, if the users
of those sites like the benefits of the added functionality, the usage
will naturally spread. The 'Replaces' header might well fall into that
category, but I cannot see how 'mvgroup' could. Please, if you have a
scenario in mind whereby a gradual introduction of 'mvgroup' could happen,
then please describe it to us.
-- 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