Danger to the Net? (Re: My thoughts about theproblems of the IETF)

[Brian E Carpenter]

Exactly why Dave Crocker and I floated the idea of a panel
of reviewers. I do intend to turn that into a draft, one of
these decades.

   Brian

John C Klensin wrote:
> 
> --On Monday, 21 April, 2003 17:13 -0700 Fred Baker
> <fred at cisco.com> wrote:
> 
> > At 06:30 PM 4/21/2003 -0400, Steve Silverman wrote:
> >> If less than 7 people have read and really understand an ID,
> >> there is a good chance, that there
> >> are significant undetected problems that may result in an AD
> >> blockage or problems down the road.
> >
> > That suggests a procedural step: any working group that sends
> > a document to the IESG should perhaps have a requirement to
> > identify 7 people that read it.
> >
> > I'm not generally much for idiocy tests of that sort, but in
> > this case one wonders...
> 
> Fred,
> 
> I don't know if seven is the right number, but I have believed
> for some time that many steps in the approval process would be
> improved if the submission process from a WG to the IESG
> included a checklist document that, in turn, included
> 
>         * An explicit statement from the WG chair that the WG
>         had reviewed the document and reached meaningful
>         consensus.
> 
>         * Checkoff statements (one at a time) that the required
>         sections were present or explanations as to why they
>         were not required.
> 
>         * Explicit signoff from several participants in the WG
>         that they had read the document, verified it against the
>         checkoff statements, and were of the opinion that the
>         document was technically competent, an adequate
>         definition, complete, and represented WG rough
>         consensus.
> 
> That "several" could reasonably be your "seven".  And the notion
> would be that, if the responsible AD, or the IESG, detected
> document problems that this process should have identified and
> didn't, we would have a list of names that could be held up for
> public ridicule and as responsible for any ensuing delays.
> 
> Pretty hard-nosed, but there really is no excuse for asking an
> AD to spend time reviewing WG output that has long-required
> sections missing, that is not intelligible, or that no one has
> really read.
> 
>       john