In Progress Bug results

[John Arbash Meinel]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

So I just went through and re-triaged all In Progress bugs. I figured I
should give a summary of what I found, and what actions I took.

When I started this the other day, we were at 67 bugs In Progress, right
now we are at 40.

I also want to say that I hope I didn't step on anyone's toes. We value
work that has been done, but if something hasn't been touched in 2
years, I don't know that we can bulk fix all of them right away. If it
motivated you to finish your patch, then great, set it back to In
Progress, and lets get it landed. If you just need help, please contact
us and we'll do what we can to get things completed.


For Canonical[1] folk, the graph over the last year is here:
https://lpstats.canonical.com/graphs/BzrProjectBugsInProgress/20100422/20110422/

Anyway, that shows a peak of about 84 in-progress bugs last year, which
is down to 60-something today, and should be about 40 tomorrow.

Also interesting is this graph, which is since January, 2011.
https://lpstats.canonical.com/graphs/BzrProjectBugsInProgress/20110101/20110422/

Which shows we've been pretty steady between 60-70 for all of this year.

1) There are a fair number of genuine things in progress. Probably 30+
of them. Mostly just the actual work queue of a few people. If you
figure 4-5 very active core devs (with 4-5 in-progress each) + 10
community works in progress, that gives you 35 In Progress, which is
quite comparable to 40. I think Martin's target was 30, which is
probably only achievable by actually getting Jelmer to land all of his
excellent work :). (And getting Martin to not set it as In Progress if
it is something "I'd like to do in the future".)

2) There were a lot of things with 2yr+ patches that have gone stale. If
it looked reasonable, I tried to pick it up. But if it didn't I:

 a) Changed In Progress back to Confirmed. If nothing has progressed in
    >6months I think it is fair to say it isn't *In Progress*.
 b) Added the tag 'patch-needswork' so that we can find these as Patch
    Pilot and poke at them if they are interesting.
 c) I think these are good cases for either casual people or the patch
    pilot. Any work that is done but not landed is essentially wasted,
    and we really do value people's work. However, some are not past
    the point where it would be less work than re-implementing from
    scratch.

3) I think it would be an interesting goal to burn down "Work In
Progress" proposals, Bugs with Patches, patch-needsfixing bugs, etc. It
is part of what Patch Pilots can do. I can't say what my personal
priority would be for it, though. Which is why I tried to move things
down the stack a bit that have stalled. We want to have useful metrics
for "this patch that someone is actively working on is stalled" versus
"this patch that someone gave up on 3 years ago is stalled". And to do
so, we need to push some of the older stuff out of the queue so that the
queue can be usefully scanned.

4) Even 30 In Progress bugs is a bit hard to just "scan". I would say
10-20 is something that could be used as a reasonable "how are things
going". But maybe just sorting by date is another way to handle it.

5) I made sure that every In Progress bug had someone who was Assignee.
I don't think it makes sense to say "This is In Progress" without saying
who is working on it. If I couldn't find anyone working on it, I just
set it back to Confirmed.

John
=:->

[1] I really wish we had this sort of graphing in a public place that
didn't require polling the database server directly, and thus security
concerns making it private.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAk2wR1MACgkQJdeBCYSNAANF8gCfUmLV40QVNLrPsEAt0teIuUTi
I1EAn3gM68HaQVE5dOcUGNdSfh/xx2Hi
=VZgp
-----END PGP SIGNATURE-----