launchpad-dev team mailing list archive

[Robert Collins <robert.collins@xxxxxxxxxxxxx>]

Week 1:
rev 11823 till 11875 there were 7 unreviewed landings - 13%.
Landings where a review would have been a net win 1.
Landings where review would have been the best place to improve things: 0

Week 2: (*) There are two landings of mine not reviewed yet. But they are tiny.
rev 11876 to 11926 with 8 unreviewed landings - 16%.
Landings where a review would have been a net win 0.
Landings where review would have been the best place to improve things: 0

Week 3: (*) There is one landings of mine not reviewed yet. But its tiny.
rev 11927 to 11956 with 7 unreviewed landings - 23%.
Landings where a review would have been a net win 1.
Landings where review would have been the best place to improve things: 0 or 1

The tracking table is at
https://dev.launchpad.net/PolicyAndProcess/OptionalReviews.

So, this concludes the (slightly extended) trial phase of this
experiment: every self reviewed landing has (or will be when jml
reviews my 3 self reviewed landings) reviewed.

I think the experiment is going well - there's only been 2 out of 22
landings where a review would have been worth it (IMNSHO) - so we're
making good tradeoff calls over 90% of the time.

>From here on in, each month we'll divide up and allocate self reviewed
landings to reviewers, one review per reviewer, to keep tabs on this.
Remember when doing this that the question is 'would a review have
been worth it', not 'would a review have made things better'.

Francis will organise the first of these on the 1st of November (which
only needs to capture things from 11960 on), and I'll organise the
second on the 3rd of January. Please remember to create a merge
proposal when self reviewing, so that this process can use that.

Thanks,
Rob