Differences between revisions 6 and 7
Revision 6 as of 2008-11-17 09:53:31
Size: 4865
Editor: 85-218-4-225
Comment: edits for final board report
Revision 7 as of 2009-09-20 23:35:11
Size: 4865
Editor: localhost
Comment: converted to 1.6 markup
No differences found!

Google Summer of Code 2008 - Final Report

/!\ See the minutes of the 2008-11-19 board meeting for the official version of this report.

The Google Summer of Code was in its third year this year. We received 186 applications, of which 31 were slotted. 22 of these projects passed both the mid-term and final evaluations. 5 of these passed the mid-term, but failed the final evaluation and a further 4 failed at the mid-term. This pass rate is below the average rate for the whole GSoC program, however, it was felt by the mentors and admins that this reflects the high standards expected of contributors to the ASF.

This year we had three administrators:

  • Bertrand Delecretaz
  • Ross Gardler
  • Luciano Resende

Having more administrators is a good thing and seemed to work well since the most experienced of these admins was able to spend some time documenting the GSoC processes on the wiki.

Selection Process

In previous years the application RankingProcess worked reasonably well. However, this year there were insufficient numbers of mentors reviewing all proposals. This resulted in some well represented projects receiving an artificially high ranking. The admins attempted to ensure that the final rankings were as fair as possible, but there can be no doubt that the RankingProcess needs to be revisited for GSoC2009 (assuming there is a GSoC 2009, of course).

This has already been discussed on the code-awards mailing list and we believe a fairer system, which does not rely on significant numbers of mentors reviewing all applications, was provisionally agreed upon. However, it should be noted that no ranking system will be perfect and there will, undoubtedly, always be someone who feels the system has failed them. We hope that the system adopted for 2009 will be good enough and encourage all interested committers to oversee the definition and execution of the process on the code-awards list.

Success Stories

A number of success stories were collected on the code-awards list, the full text of all SoC2008SuccessStories is available on wiki.apache.org, whilst some edited highlights are included below:

  • Milinda Lakmal Pathirage is now a committer for Apache ODE.
  • Both Robin Anil and Deneche Abdelhakim contributed code that has been committed to Apache Hadoop, and both did a good job of working with the community by answering questions and participating in the process.
  • Wojtek Janiszewski is now a committer on Apache Tuscany
  • Douglas Leite is now a committer on Apache Tuscany
  • Oscar Castaneda is now a committer on Apache Tuscany
  • Andrey Razumovsky became a committer on Apache Cayenne and is described as one of the most active Cayenne committers as of now.
  • Andras Belicza and Tharindu Mathew provided a useful new module and numerous bug fixes for Apache Harmony.
  • Lahiru Gunathilake got his committer rights for Apache Qpid in September.

This means that the ASF gained at least six committers in this years Google Summer of Code. The administrators would like to thank all mentors for their hard work during the Summer of Code.

Mentor Summit

/!\ Report from Luciano

Every year after Summer of Code is over, Google hosts a summit at its headquarters in Mountain View to foster sharing of mentors experience, and also to learn first hand what could be improved for next year program. Luciano Resende attended the Summit on behalf of ASF.

This year, the summit followed the unconference idea, and various group of people gather together to discuss Summer of Code and other OSS related topics. Various of the suggested themes were around how to successfully go from recruit to project delivery without having the student to disappear. Some mentors/organizations were trying to make the student selection very strict and really follow a job recruitment procedure, where others were more towards given an opportunity for the students to learn and enjoy working with open source software.

(following section omitted for the board report, more relevant to GSoC mentors IMHO)

Below are some high level suggestions I gathered from some of the sessions:

  • Generic Organization Project Selection
    • Give extra points to students that have already engaged on the community, and have openly discussed project ideas and project proposals on the public lists.
  • How to avoid disappeared students
    • Among various different opinions, and lots of disagreements, there seemed to be some consensus in two main advices to mentors :
      • Make sure your student understands that communication is not an option, it is mandatory and is reason to failure.
      • Be very clear with your student about your expectations, and I would probably advice to agree up front on what criteria will be used during student's mid-term and final evaluation.

SoC2008FinalReport (last edited 2009-09-20 23:35:11 by localhost)