Behind the big numbers on the Wikimedia code review process

Having a dashboard usually opens new paths to understand software development communities. This may be seen as the entry point that helps to understand the basics of a community. And on top of this, there may appear new questions related to those basics or to more advanced issues. This is the case of the new work we are working on with the Wikimedia community metrics analytics team: Core Reviewer and Participants.

  • Core reviewers are defined as those developers that can exercise a +2/-2 review in Gerrit. In addition to this, it is of interest for the community to remove auto merges. Although this is an undesired behaviour, that takes place, and those should be removed.
  • On the other hand, Participants in Gerrit are defined as any member leaving any type of trace in the system. In this set we can find reviews (-2,-1,+1,+2), uploads, comments and others.

It is interesting to notice that depending on the community, requirements are slightly different. In the case of the OpenStack community, there are extra requirements for the Core Reviewer definition. And this is that reviews should be found in master branch. This specific measure can be found in the OpenStack quarterly reports for each of the projects of the Foundation.

In order to have first numbers, we decided to go for a small and initial analysis through GrimoireLib. Indeed we had to add a new metric, named as Participants that can be found in the code review process metrics module and add a new type of filter to ignore if required auto merges in the system. This would be used by the Core Reviewers metric later.

Based on the numbers retrieved from the software development metrics of the Wikimedia Gerrit database, the analysis shows a total of more than 900 different people that at some point have participated in the review process. According to the following chart, there were two peaks of participants, in March 2014 and in July 2014, with around 280 participants in each of those two months.


Evolution of participants in the code review process of Wikimedia



In addition, the Core Reviewers analysis shows a continuous increase in the number of people executing actions as +2 or -2.


Evolution of core reviewers in the code review process of Wikimedia


The following steps consist of adding those two new metrics as a chart plus a list of contributors in the review process overview panel or in some of the KPI’s panels defined in the same dashboard.

The data provided in this analysis is available as ipython notebooks in the Wikimedia analysis. And through the nbviewer application: Participants and Gerrit Core Reviewers.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s