Chandler Engineering Meeting - October 12th, 2006

Today's meeting was entirely devoted to alpha4 status and trying to get a sense of when a Release Candidate could be cut.

Blockers

We decided during the Bug Council this morning to use the "Blocking0.7" flag to mark real alpha4 blockers. During the day, we checked with devs and voted for some bugs (using the "?"). We reviewed all the proposed blockers during the meeting. Results:

• 13 blockers
• 3 left as blockers "?" (to be investigated)

Those blockers need to be fixed or we need enough investigation that we know what we're leaving in there. Devs need to concentrate on those blockers first thing.

SWAG

To fix those blockers, taking into account the feedback from devs gathered during the day, we'll need a week (starting now). Several devs are on the "critical path" (i.e. several people have a week of work in front of them to absorb those known bugs).

So alpha4 will not happen by Tuesday October 17th as planned. The earliest would be next Friday.

Find Rate and Stability

The find rate concerns all of us. Bugs have been fixed at an amazing speed by the dev team but we keep running into serious bugs every day. We all feel that we need a couple of "quiet days" (i.e. 2 consecutive days without a blocker being discovered) before we can feel confident that the code reached an acceptable level of stability.

We do not know when we'll reach that point but since we are committed to providing a really dogfoodable alpha4, it seems to us that this is a minimum criteria we need to meet.

Consensus

The Alpha4 date is moved one full week to Tuesday October 24th so to allow us to fix the identified blockers. Hopefully, by that time, we will have had a 2 days quiet period. If we haven't, we will continue to slip the date till we get to this stability goal (again, that's 2 days without new identified blockers logged, that's not 2 days without any bug logged though...).

-- PhilippeBossut - 13 Oct 2006