In the last year we have organized a few workshops as one of the ways of find new people that would help us test software that runs Wikipedia. During one of the first workshops (maybe even the first one) a new contributor appeared. At the time all I knew about her was her IRC nick, rachel99. After the workshop she continued to contribute Cucumber scenarios.
Then, Wikimedia Foundation decided to join a few other free software projects in Outreach Program for Women (OPW). (My colleague Quim Gil, our OPW coordinator, wrote more about this round of OPW.) Rachel applied for the internship focused on browser test automation for VisualEditor.
In a lucky coincidence, I was in Boston for Selenium Conference 2013 a few days before the internship started, so I had a chance to meet Rachel in person.
Is the project completed?
Rachel wrote several tests for VisualEditor. At first they were in our "sandbox" code repository but before the end of the internship, we have moved the test to VisualEditor repository (you can find them in modules/ve-mw/test/browser folder).
Jenkins is running tests that she wrote at least twice a day. We plan to run the test more frequently at the future (before a commit is merged into master branch), but we are not there yet.
What have we learned?
Remote pair programming works. We have used Google Hangouts, but I have previously used Skype and iChat and all of them worked just fine. I have paired with Rachel in one hour sessions several times a week, sometimes almost every day. Chris (my colleague and Rachel's other mentor) paired with her probably as much as I did, maybe even more.
Every intern should have two mentors. The internship was over the summer when people take vacations, so if a mentor goes to vacation for a few days (or weeks), the intern could get stuck. Also, sometimes a mentor could be really busy or traveling. Having two mentors makes that situation less likely.