Hi Eliot,
Just on processy points... ;-)
On 09/07/14 09:48, Eliot Lear wrote:
Given the use of normative language, you may also wish to
consider making this document a BCP and reviewing it in that light.
We (the IETF:-) agree that one can use 2119 language in informational
RFCs, as we do it all the time:-) In this case, I think myself that
the 2119 language isn't really needed but I raised that on the saag
list and nobody seemed to have a problem with it as Viktor has it,
and I don't feel strongly about it. (I only raised it in the hope of
not having to talk about 2119 during IETF LC:-)
On the BCP thing, I think that would be a mistake, for now, for this
document. This is really just defining and describing OS so that other
protocols can more easily use that technique, should they want to,
without having to argue terminology and re-define the basic concept
over and over with minor and subtle variations being likely each time.
So for example, in the httpbis wg case this document would mean the
wg wouldn't have had to have had the "what is OS and is it ever a
good thing?" discussion, but they still would've had to figure out
whether and how to use OS in e.g. HTTP/2.0. (As has happened, and
is getting close to done I think.)
Later, I think we may want to provide guidance about when to use OS
and when not, and that's when I figure the BCP part of the OS story
would get done. For that I think we need to learn from what's
happened in e.g. the httpbis wg and elsewhere, so I figure we're
not there yet. And even though we do IMO have a really good success
story for OS with recent deployments of STARTTLS for MTA-MTA SMTP,
it'll be interesting to see if the non-authenticated cases there
transition towards authenticated endpoints or not over time so we
might be better off waiting a while to find out stuff like that
before writing BCP text.
Cheers,
S.