The Oregon State University Open Source Lab (OSL) participated for the first time in the Google Code-in (GCI) contest, and we're pleased to report that the results of the contest were incredibly valuable for us. If you haven't heard about Google Code-in, it's a contest run by Google's Open Source Team that aims to introduce pre-university students to all aspects of open source software development. (For more details, check out the excellent summary on the Oregonian's Higher Education Blog.) We had eight students actively working with us in GCI, from as far afield as Australia and Poland. Of these eight students, two have let us know they plan to keep working on projects with the OSL and one even let us know that he's hoping to attend Oregon State University so he can learn even more about open source and help out at the lab.

We had a total of 42 tasks completed, all improvements to Ganeti Web Manager, a homegrown project of the OSL. For those who aren't familiar with the project, it's a Django based web application that help administrators better manage their Ganeti-based clusters. Our students completed a total of 42 tasks: 23 coding tasks, 18 tasks to improve Ganeti Web Manager's user interface, and one outreach task (logo design). We were especially excited that our students found several unknown bugs in our code base and proceeded to send us fixes for them. You can take a look at all of the tasks proposed by the Lab and all those completed by our students on the contest website.

As a result of participating in Google Code-in, several major features were implemented in Ganeti Web Manager, including:

• HTML5 VNC Support
• Creating a Status Dashboard for Administrators and Users
• Creating an Object Change Log

Of all these major features, creating the status dashboard was the most difficult of our tasks. We required a very concise layout for the feature and the student working on the task was not a native English speaker, but he did an outstanding job. In fact, if you'd like to get to knw the student who worked on the dashboard, we've already published an interview with Piotr for your reading pleasure.

We also noted down some observations about the contest that we'd like to share in case they are useful to the community in planning for any future GCIs. We had about 10 students sign up for difficult tasks and simply never submit any work for them, though in two cases the students submitted an initial set of work and then did not follow up after receiving feedback. We're wondering if other folks had this experience and what they did to mitigate it. We certainly did our best to be available as much as possible either via email or on IRC. We're very happy with what our students accomplished, though, so perhaps its just that those students who abandoned the tasks found they weren't the right fit for them and didn't know to withdraw their claim.

We've found that GCI was far more work for our mentors than our participation in Google Summer of Code, though in some ways it also much more productive. Our mentors found it easier to select smaller, bit-sized projects for students to work on, though that also meant a lot more work to review their output in a short period of time. One of our mentors notes that the value of GCI is in that same rapid pace, though; if a student discovers that a task is not to his liking or the mentor realizes she may have sent the student off in the wrong direction, it's a quick fix to send the student along to the next task or to a different mentoring organization that may better match the student's interest. With a 3-month long coding project, it's less easy to be agile in deciding when to rework the project or when it is best to simply not complete it.

Overall, we were incredibly excited to be a part of GCI and extremely pleased with our results. We certainly hope to continue mentoring for the contest, assuming Google chooses to run it once again. Many thanks to all of our students for their great work and many thanks to Google for doing awesome work to promote student involvement in open source!