RealTime IT News

Was Google's Summer of Code a Boon or Bust?

Last summer, Google sponsored 410 open source projects spanning 41 different organizations in a development initiative known as the Summer of Code.

How successful were those projects and how many of those Summer of Code 2005 projects still survive?

The number may well be less than half.

Chris DiBona, open source program manager at Google, wrote in a blog post that "something like 30 percent of the students stuck with their groups post SoC [Summer of Code]."

But DiBona noted that the SoC isn't only about getting coders to stick around in the projects after the program.

"We also were trying to introduce students to the open source development process, keep students working on coding over the summer than take a non-cs related job and get some new code written," DiBona wrote in an e-mail to internetnews.com.

Some projects fared worse than the 30 percent survival ratio, while some fared much better.

Mozilla Developer Gervase Markham noted that, of the 10 Google-sponsored Mozilla projects, not one had any sign of work continuing past the end of the SoC deadline.

"Not one. No development-related mailing list traffic, no releases, nothing," Markham blogged. "It's as if the people vanished off the face of the earth on September 2nd."

There were exceptions.

The GAIM instant messaging project had a great deal of success. Nearly all of the projects have found their way into the version 2.0 of the IM client, currently under development.

Sponsored projects at the Jabber Software Foundation fared slightly above the average.

According to Peter St. Andre, the executive director of the foundation, out of the 10 sponsored projects, only eight were completed. Of those eight, only three or four are active.

The KDE project, a graphical desktop environment for Linux and Unix, had the second-highest number of sponsored projects, with 24. KDE fared somewhat below the average.

"Of the 24 projects, only one saw active development in the meantime," Thiago Macieira, KDE core developer, told internetnews.com.

That doesn't mean that the code was wasted, Macieira said. It only means that only one SoC project code was been developed by the student after the end of SoC.

Jabber's St. Andre noted that he wasn't surprised by the low survival rate for the summer projects.

"Given that SourceForge and other open-source repositories are filled with abandoned projects, I can't say that I'm surprised," St. Andre told internetnews.com.

St. Andre explained that the purpose of the Summer of Code is to introduce students to open source development, and not necessarily to execute serious projects.

"So as long as students are introduced to open source and learn something from their mentors, I think the experience will have been worthwhile," St. Andre said.

"I realize that some people would like to pin greater hopes on the SoC than is warranted, but IMHO that's a problem with their expectations, not the program itself."

Google's Summer of Code for 2006 is just about to get underway, with participant projects expected to be named next week.

KDE's Macieira noted a few measures his project is taking to help ensure a greater degree of project survival this year than last.

"We're taking more time to select the projects this year," Macieira said. "This year, we're looking more into how likely those applications are to accomplish what they propose and how the project would benefit from it as a whole.

"So we may pass up some very interesting applications in favor of those that we feel the student to be more conservative and within his grasp."