Brian Wolff bawolff@gmail.com wrote:
Alright, let's not attempt to mix new contributors and code review, then. Thank you for your good arguments.
if the issue at hand is to review the ~ 2500 "code re- views waiting for reviewer", if every one of the ~ 100 WMF employees reviews one (additional) changeset per workday, the task will be done by the end of this month.
As far as I can see, there are not 100 WMF employees capable of reviewing changesets, even less of merging them. Not only this, expertise and +2 rights must be seen from the point of view of repositories, because not every reviewer can review every patch. Once we solve "Bug 63533 - Gerrit metrics about open changesets should ignore -1s", we will be able to see what kind of problem we have, and that will help us decide what needs to be done.
Indeed. as it stands, there are only 78 people who have ever +2'd even a single changeset in mediawiki/core, of which only 53 are staff. So if 100 WMF employees reviewed a change set today, for roughly half of them it would be the first time they had ever done so (I've excluded extensions from this analysis. Also I counted quickly, don't use those numbers for anything important).
Of course, by all means if you're a WMF employee who is knowledgeable enough to review something (Or if you are a volunteer that has +2 rights and knowledgeable enough), there's no better time to start reviewing things then today :)
This implies that the changesets waiting for a reviewer only need someone to +2 them in the condition that they are in at the moment. I'd guess (no data to back it up! :-)) this is not correct for the majority of changesets that will need several iterations to incorporate (valid) suggestions from the review.
Tim