Bug Council notes for 0.6 release

Criteria: dogfood calendar, PerformanceProject

0.6 bug table

Release Estimation

This is based on triaging 10% of the bugs currently targeted to be fixed for 0.6 - consider this estimate to be accurate with 10% probability

522 bugs fixed in the last 90 days by 24 people. In other words, on average we fixed 1.7 bugs per person per week, but that is counting interns and non-developers, some of whom only fixed a bug or two during the 90 days. Applications team has had a fix rate of 5 bugs per developer per week (I seem to recall from pbossut's figures).

735 bugs were filed in the last 90 days. In other words, we get 8 new bugs per day or 57 new bugs per week. 69 duplicate (9.4%), 356 fixed (48%), 8 invalid (1.1%), 21 wontfix (2.9%), 28 worksforme (3.8%), 253 are still unresolved (34%). Of the unresolved, 209 are either untargeted or targeted to be fixed within 0.6 release cycle. This could mean that of the incoming bugs, worksforme+fix+targeted for 0.6 accounts for 593 bugs, or 80%.

• Currently we have roughly 23% P1-P2 bugs
• Currently we have roughly 59% P3 bugs
• Currently we have roughly 18% P4-P5 bugs

Based on bug triage of alecf's triage, we can triage 23% or the bugs away in the first round. We currently have 321 bugs targeted for 0.6 release, so with first triage that should drop down to 247.

Situation after:

• 1 week: 247+57*.8-50=242 remaining (effectively reduce remaining bugs by 5 per week)
• 2 week: 237 remaining (these seem to be off: 181 P3, 60 P4-P5), 100 fixed
• 3 week: 232 remaining
• 4 week: 227 remaining (these seem to be off: 157 P3, 76 P4-P5), 200 fixed
• 5 week: 222 remaining
• 6 week: 217 remaining (these seem to be off: 133 P3, 92 P4-P5)
• 7 week: 212 remaining (these seem to be off: 94 P3, 100 P4-P5), 350 fixed - shouldn't really do any work here
  • 7 week if no bug fixing: 258 remaining (these seem to be off: 11 P1-P2, 121 P3, 108 P4-P5)

assuming no further triage, and assuming we fix 5 bugs per person and we have 10 persons resolving bugs. Also assume new bugs come in with the same distribution of priority (11 P1-P2, 27 P3, 8 P4-P5).

So given our target of 7 weeks for winding down for release, it would seem we must forget all P4-P5 bugs and ship with something like 150 P3 bugs (assume stopping all development after week 6), and 11 newly discovered P1-P2 bugs (although I do think P1-P2 rate is going to go down the further we go). Or if we want to have as few P3 bugs as possible, optimal point would seem to be to develop 10 weeks, with 27 newly discovered the week after development stopped.

Statistics Update

After triaging the bugs of 4 of the most doomed engineers, we've been scrubbing about 51% of the bugs targeted for 0.6 Release. If that were to continue with that rate, we would scrub about 75 more bugs. If that was done today (10/11) we would have 166 bugs in the 0.6 release + 12 in 0.5.06 + 16 untriaged.

Statistics for last 30 days

last 30 days

Regression rate about 8%.

---

Bug Council 9/27/2005 Agenda

• Triage alecf, john 0.6 bugs

Minutes

• Only managed to do alecf's bugs (but did all of them). Got from 35 to 27.
• It seems like we can triage about 20 bugs an hour. We should spend close to 5 hours a week for the next two weeks in order to go through the initial triage.

Bug Council 10/2/2005 Agenda

• Triage john's bugs (39).

Minutes:

• Got down to 22 bugs. 11 left to triage. Continue from Bug:3945.

Bug Council 10/4

Minutes:

• Got from john's 24 remaining to 20, all done.

Bug Council Agenda for 10/7

• davids bugs (40)

Minutes:

• Got down to 13!

Bug Council 10/10 Agenda

• bear (26 bugs)

Minutes:

• Got down to 8 bugs!

Bug Council 10/11 Agenda

• 0.5.06 bugs (31)

Minutes:

• Triaged all. Remaining in m6: 18.

Bug Council 10/13 Agenda

• donn (19)

Minutes:

• Got down to 17.

Bug Council 10/17 Agenda

• jed (22)

Minutes:

• Down to 11.

Bug Council 10/20 Agenda

• stearns (12)
• jeffrey (12)
• heikki (12)

Minutes:

• stearns down to 11.
• jeffrey down to 11.
• heikki down to 7.

111 bugs to go

Bug Council 10/21 Agenda

• morgen (22)
• twl (7)
• pje (3)
• pbossut (3)
• sheila (22)

Minutes:

• morgen down to 20, but he has fixes to many.
• twl down to 5.
• pje still 3.
• pbossut down to 1.
• sheila down to 20

52 bugs to go

We have scrubbed 36.5% of the bugs.

Bug Council 10/24 Agenda

• capps (20)
• aparna (3)
• mimi (7)
• vajda (4)

Minutes:

• capps down to 10
• aparna still 3
• mimi down to 4
• vajda still 4
• 8 untargeted triaged; 5 to 0.6
• 20 bugs to go

Bug Council 10/25 Agenda

• grant (9)
• bkirsch (9)

Minutes:

• grant still at 9, but there are several fixes already
• bkirsch down to 8, but lots of P4s
• All done, down to 208 bugs targeted for 0.6

Bug Council 11/8 Minutes

• Decided to release 0.5.07 on schedule
• Bug Council to start marking bugs with blocking0.6+ flag that we would hold the release for
• Bug Council start to triage with more strict criteria the incoming bugs as well as the bugs already targeted for 0.6
• Went through P1 bugs and added blocking0.6+ flag where appropriate.

Bug Council 11/9 Minutes

• Went through P2-P3 bugs and added blocking0.6+ flag where appropriate. Only 7 P4 bugs to triage.
• 140 0.6 bugs
• 33 0.6 blockers