Quim Gil qgil@wikimedia.org wrote:
In relation to our discussion about code review metrics at http://korma.wmflabs.org/browser/gerrit_review_queue.html
I just learned that OpenStack has a policy to abandon automatically reviews sitting in -1 during more than one week:
Maybe one week is too tight for the reality of our project, but what about 2-4 weeks?
Brad and others said last week that they were not interested in code review queue metrics mixing pending -1 reviews. Maybe the solution is not to tweak the metrics, but to effectively abandon those changesets automatically?
No. First of all, this would give anyone who has -1 and can click some buttons the power to abandon changes or at least whip a contributor into action: "Fix this /now/, or else!" I don't think this would be a healthy atmosphere for devel- opment. From my limited view of development at OpenStack, it appears that doesn't force contributors to produce +2able changes in a jiffy, but just give up.
Second, changes with -1 appear in the Gerrit UI by default, while abandoned changes do not. So all the work that was done is effectively lost when finally someone new comes along and wants to tackle a problem that has been approached before.
Third, this sends out the message: "We welcome you as a con- tributor! A patch, how awesome! Oh, sitting more than four weeks? Then go away and only come back when your code is ready because you are messing up our statistics!"
I'm all for abandoning changes when the author doesn't react and the patch doesn't apply anymore (not in a technical sense, but the patch's concept cannot be rebased to the cur- rent HEAD). But forcing work on many just so that a metric can be easier calculated by one is putting the burden on the wrong side.
Tim