Stringency (RE: 12 problems)

[James Kempf]

Dividing the delay by the number of RFCs gives an approximate estimate of
productivity, how many months it takes to approve an RFC. This number does,
indeed, show a bit of a decline between 1995 and 2000, with the steepest decline
happening between 1999 and 2000. The productivity declined between 1993 and
1994, but I assume that was due to the sharp jump in the number of RFCs between
the two years, it recovered again in 1995 but then started to slide:

1993    0.177
1994    0.105
1995    0.172
1997    0.139
1998    0.133
1999    0.126
2000    0.010

Might be interesting to weigh these with the number of participants in working
groups, to see if that also had an effect.

            jak

----- Original Message -----
From: "Henning Schulzrinne" <hgs at cs.columbia.edu>
To: "Harald Tveit Alvestrand" <harald at alvestrand.no>
Cc: <john.loughney at nokia.com>; <problem-statement at alvestrand.no>
Sent: Thursday, January 16, 2003 3:18 AM
Subject: Re: Stringency (RE: 12 problems)


> http://www.cs.columbia.edu/~hgs/internet/IMR has the data and the .tcl
> scripts. The measurements I reported were:
>
> year number of RFCs sampled  avg. delay (months)
> 1993 31                      5.5
> 1994 80                      8.4
> 1995 61                     10.5
> 1996 96                     12.4
> 1997 103                    14.4
> 1998 147                    19.6
> 1999 157                    19.8
> 2000 158                    24.2
>
> Things change a bit on the tail end if we ignore all RFCs with duplicate
> titles, where association with I-Ds is likely to be iffy:
>
> 1993 27 5.6
> 1994 73 8.0
> 1995 50 8.8
> 1996 90 12.2
> 1997 98 12.9
> 1998 123 15.0
> 1999 147 18.8
> 2000 139 19.1
>
> http://www.cs.columbia.edu/~hgs/internet/id-rfcyear.png
>    Number of RFCs and I-Ds published each year, based on IMR reports from
> the year 1991 and later
>
> http://www.cs.columbia.edu/~hgs/internet/rfcyear.png
>    Number of RFCs published, also ratio to IETF attendance
> (Ignore the year 2000 dip, as that's based on year-to-date statistics)
>
> Harald Tveit Alvestrand wrote:
> >
> >
> > --On torsdag, januar 16, 2003 06:28:32 +0200 john.loughney at nokia.com wrote:
> >
> >> Harald,
> >>
> >> Any stats of how long I-Ds take from WG Last Call completed to being
> >> shipped to the RFC editor?  If this was examined over time, we may have a
> >> better metric to see about IESG stringency ... If the this time has been
> >> ballooning - especially in relation to other parts of a document's
> >> lifetime, then we should assume there is some problem.  If it is not
> >> dramatically increasing, then perhaps stringency would not be a problem.
> >
> >
> > Unfortunately the stats in the ID-tracker haven't been around long
> > enough to make sensible trend lines.
> > A few years back, Henning Schultzerinne made scripts chewing through the
> > Internet Monthly Report (http://www.ietf.org/IMR/) to generate stats on
> > the time between -00 version publication, IETF-wide Last Call and RFC
> > publication - that's the source of the frequently quoted statistic that
> > "time has gone from 6 months to 2 years".
> >
> > I've CCed him on this message; if he still has the scripts, and is
> > willing to share, we could get at least those numbers back again....
> >
> >               Harald
>
>