Re: possibly entertaining statistics

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Thomson, Martin wrote:
On 2011-09-07 at 20:13:34, Jari Arkko wrote:
RFC publication speed (time from draft-smith-00 to RFC, measured in
millibits/s):
http://www.arkko.com/tools/lifecycle/sizematters.html

Jari, throwing bait to systems modeling engineers is not right! :)


This isn't particularly empirical, but it would seem that of the faster drafts, having the three characters 'bis' in the title helps.

Curious about that, I imported the data into a SQL table and see 63% of the total 329 bis docs reaching RFC status.

Here are the totals of BIS docs that have a RFCDATE = 0000-00-00 (field #16) vs the current state (field #22) which I took to mean it never got an RFC status:

+----------- RFCDATE = 0000-00-00 ------------+
| TOTAL| STATUS                               |
|------+--------------------------------------+
| 1    | AD is watching - External Party      |
| 1    | Publication Requested                |
| 1    | DNP-waiting for AD note              |
| 2    | AD is watching - AD Followup         |
| 2    | Waiting for AD Go-Ahead              |
| 3    | IESG Evaluation                      |
| 3    | IESG Evaluation - Revised ID Needed  |
| 3    | IESG Evaluation - AD Followup        |
| 6    | AD is watching                       |
| 7    | RFC Published                        |
| 10   | RFC Ed Queue                         |
| 28   | Dead                                 |
| 54   | ID Exists                            |
+---------------------------------------------+

That 7 count for "RFC Published" seems off, of course, probably reflects docs that did not get rfc date recorded in the tracker.

Overall, I think this is pretty good - we are careful! :)

I was curious to see how the # of revs correlated to the days from initial I-D to RFC publication date.

Total     :  8391
RFC Status:  5525 or 66%

While there were scattered data points # of revs vs the days to RFC publications, the pattern was faster RFC publications correlated to less revs were necessary.

Of course, this is only significant to show the more revisions an I-D has reflected an vetting and contention process (thats good) and the longer it took to progress to RFC status.

Then again, there were drafts that took years with 0 to 1 revisions.

For example 583 of the drafts had 0 revisions and ranged from 44 days to less than 10 years to get an RFC status and 714 drafts with 1 revision had a range of 98 days to over 10 years to get an RFC status.

To me, its puzzling why many of these docs with [0,1] revs took so long as I would think that the less revisions one needs the closer it is to the "ideal protocol" it wanted, but then again, it may reflect the lack of WG vetting involved, so in the that vain, the system has worked.

Anyway, interesting data.

--
HLS

_______________________________________________
Ietf mailing list
Ietf@xxxxxxxx
https://www.ietf.org/mailman/listinfo/ietf


[Index of Archives]     [IETF Annoucements]     [IETF]     [IP Storage]     [Yosemite News]     [Linux SCTP]     [Linux Newbies]     [Fedora Users]