time-to-approve etc. [Re: Trusting the IESG to manage thereform process (was:Re:Doingthe Right Things?)]

[Pekka Savola]

On Mon, 9 Jun 2003, Harald Tveit Alvestrand wrote:
> --On mandag, juni 09, 2003 15:06:50 +0300 Pekka Savola <pekkas at netcore.fi> 
> wrote:
> 
> >> As I noted at the time, the variance of time-to-approve is quite
> >> large, even for documents that are approved by the IETF as-is. That's
> >> not completely inconsistent with document quality, but if the
> >> documents that are taking a long time to clear (>100 days) are really
> >> that low quality, it's somewhat surprising that they aren't sent back
> >> for revision.
> >
> > It might be interesting to note a few more fine-grained metrics for
> > determining the document quality, e.g.:
> >
> >  - the number of DISCUSS votes raised in the IESG
> 
> this could possibly serve as a metric of the number of *different* concerns 
> raised by the IESG, but because of the IESG's common practice of saying "no 
> further objections" when everyone agrees that the important problems 
> they've seen have already have been flagged, it's not a measure of the 
> percieved depth of the problems found.

I agree; it's more important to measure if there is a discuss vote or not
(if not, I guess the doc approval should be a very quick process :-), but
at least in my perception, several very problematic documents have had at
least 4-6 DISCUSS'es from different people.. so I guess it could yield
*some* information, but with some standard disclaimers..  

I think you could say that the more DISCUSSes a document has, the more
potential it has to be a document requiring much more work.  However, this
does not say anything about documents with one or few DISCUSSes: those may
or may not be higher quality documents (depends on what kinds of comments
will surface from different IESG members etc.).
 
> >  - documents which were approved as-is, but an RFC editor note was added
> > by the IESG (causing some delay)
> 
> I don't think this adds delay, except for the cases where:
> - the IESG takes time to write up the RFC Editor note

> - the author or WG objects to the content of the note
> The first happens a few times, but is usually (in my recent experience) on 
> the order of a week. Have people seen the second happen much?

I was thinking of the first.  But the issue is that I don't know how long
it takes to write these.  It's typically given as an action point, so I
think at least two weeks is a norm.  Tracking this could be useful.

-- 
Pekka Savola                 "You each name yourselves king, yet the
Netcore Oy                    kingdom bleeds."
Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings