ietf
[Top] [All Lists]

Re: call for ideas: tail-heavy IETF process

2013-05-01 16:47:34

On May 1, 2013, at 5:00 PM 5/1/13, Scott Brim <swb(_at_)internet2(_dot_)edu> 
wrote:

A draft does get cross-area review, at least once, often more than once.
Some drafts in some WGs get it earlier than others, by explicit
invitation.  Others don't get it until the latest they can, when they go
to last call ... but a process point for cross-area review during WG
handling, already exists.  It's just optional.  I did like SIRS, and the
equivalent of SIRS exists -- in the informal relationships that are all
through the IETF.

You can make the option more explicit, you could even make it mandatory,
but doing so won't speed up draft processing at all and will probably
make it slower.  One of the IETF's problems is not the problem under
discussion but unconscious assumptions about the standards process.
People think of last call as near the end of the process.  It's time to
abandon that outdated view and recognize that so-called "last" call is
not as close to the end as it used to be, that it's a bit closer to the
middle ... and that that's really okay.  Really.  Some other SDOs have
their process structured that way.  Let's rename "last call" to
something like "IETF review" and stop giving people the wrong
expectations.  Review outside the WG is vital, can be done repeatedly,
and must be done by the whole IETF at least once.

At the same time, we have to emphasize that "IETF review" means "a significant 
group of people have to look at this document really, really carefully to make 
sure it's right and report the results of their review".  It doesn't mean 
"well, it looks OK and the IESG is going to give a really, really careful 
review anyway so I don't need to worry about it".

Perhaps some of the directorate reviews ought to happen much earlier in the 
process?


Scott

- Ralph