While it has been shared that Debian is not a partner organization this time around, there hasn’t much information in the mainstream media about why Debian isn’t taking part in this year’s GSoC. Below are some of the reasons why it happened:
Debian has been continuously taking part in GSoC since summer of 2006 with the highest number of proposals and mentors in last year’s 2016 GSoC with 16 projects to show. With Stretch’s release also looming around, it probably made much more sense to fix all the RC-bugs rather than be part of GSoC.
This is good as Debian gets some breathing space to release Debian and also have time to think of new ideas which will make Debian awesome.
For more details, you could look at Nicolas Dandrimont aka olasd mail here.
Co/Dual mentor policy for GSoC
Debian for a long time had a single mentor policy. The disadvantage of that policy is that if a mentor falls ill, or loses her(is) job or anything untoward happens in the life of the mentor, the mentee suffers for no fault of her/his. This shows the mentee as well as the project in a bad light. Like several organizations, Debian too has decided to encourage and have 2 mentors for any of sub-projects under GSoC.
While it may have some initial hiccups as it had this year, the idea is having 2 people who know code-base of a sub-project intimately makes sure that the sub-project does not suffer in case of any disturbance in Debian mentor/Debian sub-project member’s life. It also nudges a bit all the single-person teams (and there are a few of them in Debian) to train and have a co-mentor. This should strengthen the project quite a bit as the Debian bench capacity will increase, increasing flexibility and dreaming and chasing more innovative ideas.
This was taken from the wording in Nicholas’s mail of call of participation last month as can be seen here.
See the wording “As additions this year, we ask that all projects be supervised by at least two mentors,” – Nicholas Dandrimont – DDA – Just to make sure, I did get it confirmed by him yesterday night itself.
While it does not affect Debian’s participation in the GSoC project, there were and will probably be a bit of decrease both in quantity and quality in GSoC applications this year and subsequent years due to Google’s change of policy in student stipends.
If you look into 2006 project, they paid $4500 while 2016’s stipend was at $5500. To make the amounts less than half when the rate of inflation is up is not a motivating factor and students may not be that excited for it.
To add to that, many of the GSoC students chose part of the payment received to have invaluable face-time with their mentors and network with other like-minded people at various conferences around the world. With the reduced payouts, we may see reduced inclusiveness and participation from developing countries in most technical conferences around the world.
Lastly, does this mean that Google is running out of money for the GSoC program? Many of the projects that Google helped and got good publicity from are in a better shape than probably when they started.
If GSoC were to close in a year or two, while it may be a shock, Debian would be able to scale up the Outreachy program. I wish Google shares its budget for the GSoC program every year in the interest of transparency as well quantifying the value created by Google in terms of the number of payouts to different cultures, people etc. Just my 2 cents.