My thoughts about the problems of the IETF

[James Kempf]

> One possible idea is an "approach review" or "architecture review"
> earlier in the process. Say, at the time a draft becomes a WG item
> there needs to be an "protocol overview" section which outlines the
> main technical ideas in the protocol. This section needs to be
signed
> off by a review group. This group could be e.g. IESG or a set of at
least
> one WG chair from all areas. As a part of the review, the WG could
> get useful feedback, such as "Ok, you can go ahead with this, but
you
> must add a congestion control mechanism into your protocol". This
> would be a more constructive approach than finding a lot of problems
> at the end of the process.
>

I have a prototype draft proposing an architectural review process on
my hard disk waiting for the solution phase to commence.

The basic idea is to give WG chairs and shepherding AD a semi-formal
option of convening an in-person architectural review at a WG meeting.
The review would be organized by IAB in consultation with the WG
chairs and shepherding AD, but the review board could contain any
"outsiders", i.e. those who are technically knowledgable but not
directly involved in the WGs efforts. The option would be at the WG
chairs' and shepherding AD's discretion, i.e. not required and
probably not done very often. The results of the review would be input
to the shepherding AD, broader IESG, WG chairs, and WG to guide
further development, that is, these groups could dispose of the advice
as they see fit. Making it semi-formal, i.e. documented in an RFC,
would provide some structure around a process that now is fairly
unstructured.

            jak