Friday, August 17, 2012

Yellow Squad Weekly Topics: August 17

Summary:
  • bac/gary_poster: re-reconsidering required pre-implementation calls
  • gary_poster: postpone reviewing the collaboration feedback experiment?

[introduction] [project report] (no tricks this week) [topics]

Attendance: bac, benji, gary_poster.
Apologies: frankban.

bac/gary_poster: re-reconsidering required pre-implementation calls

We had a kanban card take more that 24 hours in an active lane, and by our rules that means we look at it on our retrospective call to see if it suggests a process we could improve.

In this case, bac reported that he started down the wrong path on the coding for this card. When we recognized that the card was taking too long during our daily kanban review, bac and gary_poster discussed the branch. They agreed on a simpler and arguably more correct approach.

What can we improve? bac felt that this was not indicative of a necessary process change, but of a temporary lapse in focus. gary_poster hesitantly proposed requiring regular pre-implementation calls, as briefly discussed during last week's call. No-one, gary_poster included, was particularly excited about this idea.

Silence reigned.

In the absence of deeper analysis, we left the topic. When something like this recurs, we will work harder to determine root cause and remediation, possibly using a "five whys" approach.

gary_poster: postpone reviewing the collaboration feedback experiment?

Today was the day we were supposed to review the personal feedback loop experiment.

Because of interviews and vacation, we haven't had a normal environment for a test. Collaboration has been lower than usual because we've been less available than usual.

Do we have enough information to stop the test, deeming the approach a failure or a success? Do we want to continue the test?

benji and bac both felt that we shouldn't stop the test as a failure yet. bac pointed to the spreadsheet we have now (set up by bac using Google Forms) as a clear, accurate and valuable way of seeing that in fact we are not collaborating all that well lately.

benji wondered if we could take this time to evaluate whether we could improve anything about the test.

In that vein, bac noted that when benji had collaborated with someone outside of the squad in order to contribute a pgbouncer fix, he had entered it into the form as a collaboration with himself, because that was the closest option that was available. gary_poster noted that this was equivalent to the "external communication" metric that we had tried to identify separately. We agreed that we could add an "other" collaboration partner to our form and have this represent our external communication data gathering.

We had no other comments, and agreed to revisit the experiment at the end of September.

No comments: