Blender NOT accepted into GSoC 2015

http://lists.blender.org/pipermail/bf-committers/2015-March/045026.html

Quite a surprise. Or perhaps not given the completion rate over the past few years. Interesting nonetheless.

Could be that recent years saw the BF passing all of the students, but few of the projects actually make it into Blender.

Or, one could hope that the BF as an organization is now too big and too well funded to need the assistance of the GSoC program (with services like the Blender Cloud and an ever growing paid development team).

I can understand Google’s point of view, you give money to students and then some of them they don’t really payoff the value invested in them by not providing the results they’ve been expecting. I would say about 10% of the code write during a GSOC it gets finally in the master.

That’s possible, as even at the end of the gsoc 2014 Ton pointed that some students while they passed didn’t delivered the expected result of their gsoc paid work
http://lists.blender.org/pipermail/soc-2014-dev/2014-August/000225.html

Arg… I had a strong premonition (not in a mystical meaning). Time for an analysis :slight_smile:

Yeah, gsoc doesn’t like it when student projects regularly don’t get into master or release. It’s what plagues gimp’s gsoc participation.

Which was why Ton was hopefully wanting to make it more successful this year by making the projects more focused on specific items.

The issue with the last few years was because the BF accepted these large-scope projects from students (like retopology), that naturally has little chance of being ready for master when you take into account that the student may be touching the Blender code for the first time).

Now someone on the mailing list suggested if the BF can do their own thing, but that would have a cost of 60K and the BF doesn’t have that kind of money. However, Ton has just mentioned that there’s enough funding to slot in a few additional projects in addition to the current ones (though what they are won’t be known until later).

Great news! I hated to see all that money wasted.

I proposed on the mailing list, doing a micro Kickstarter for features we like, and that way we don’t need gsoc this year.

Ton can even control what proposals are on the board to be funded. Keeping his BDFL model.

Disappointing, but not terribly surprising. Given the reasons behind GSoC, it is not hard to see why Blender failed to be accepted. How much of the code made it into master/trunk? How many of the students stuck around to become regular contributors? For the money that went into GSoC for Blender, how many of the goals for that money were achieved?

Before people wave this off as an attack on the Blender Foundation or Ton, let me clarify that I’m not laying this at their/his feet. There were many reasons that most the GSoC projects failed to live up to expectations / requirements, some could be controlled by Ton/BF but most could not.

I think, to get back into GSoC next year, the Blender Foundation will need to demonstrate a desire and capability to bring on new developers contributing to Blender. The Kickstarter* idea has merit but any such campaign will need to take into account that the last one failed to reach targets set and that BFI/Ton intend to run another one to continue the Gooseberry work once their “trailer” is finished.

The problem is that a full 30 minutes before you sent your email, the question had already been addressed. The Blender Development Fund already exists. And if you want to fund development of your own pet feature, that’s always been possible.

I mean Ton chooses some things that he want’s but it’s not in the budget,

Put up “micro” funds for those projects, themselves.

Not as many people will fund a non BF kickstart.

Like

Specific thing 1 - Our cost estimate - Time frame from launch day estimation

this way, he can speed up blender development, and people feel more in control of the software they use.

Dollars = votes.

Well, who is actually setting the goals? Is it Ton/BF, or the students?

Needless to say, if you set highly ambitious goals, for largely inexperienced students, failure is the most likely result.

Perhaps this is ok from BF’s perspective, because it’s part of an “aim high” strategy, but for Google, it would probably seem like they’re just throwing money into a black hole.

I’m not saying that BF is blameless in the outcome. They have their share as well. Thing is, they are not the only ones that share responsibility for the outcome. It would be easy to point both barrels at the BF & fire but the students (and expectations by community) share some of the blame as well.

In case it was unclear, I completely understand Google’s POV on this. Blender is pretty much the Open-Source project for it’s niche. Wings3D, Art of Illusion, etc do have some value but Blender is the 3D graphics application in the open-source world. However, no business keeps throwing money at a project when most of the money is wasted.

Ton has had feedback on the matter. Copied from the bf-committers mailing list:

Hi,
Below Google’s feedback.

Unfortunately, as usual we had to make a lot of hard decisions this year as we always do, and your org was one of them. We also felt that your ideas page could have used some work this year. We would like ideas pages in our preferred format http://en.flossmanuals.net/GSoCMentoring/making-your-ideas-page/ and to be a bit more fleshed out than yours was this year. I hope you’ll consider applying again in a future year!

Cheers,
Carol

Our ideas page structure was copied from 2011, 2012, 2013 and 2014.
We never heard that it was not formatted correctly… OK, now we do! :slight_smile:

Here’s a good page outlining Google’s process for selection:
http://en.flossmanuals.net/GSoCMentoring/org-application/

They try to make it clear that no org can expect to be accepted every year.
That’s only fair. No hard feelings here, GSoC was always a gift, not a right.

On the GSoC mentor-org mailing list someone mentioned (two years ago) that their org skipped a GSoC once, just because they were feeling that effect was wearing off - mentors not sufficiently being motivated either. To some extent I think we have a similar effect going on… GSoC was turning into a burden as much as an opportunity for our organization. For the hardest working contributors it meant every year that their hands were tied during the summer. They now are free for other tasks - something that can also bring good things for everyone.

As for me, instead of GSoC coordination I will spend time with the module teams on streamlining our organization to make it easier for more developers to come on board. Bug tracker duties and patch reviewing seem to eat up our energy and lust-for-coding too much as well. Too few people do too much now…

HAR! Thanks Mr. Dragon. I needed a laugh.

So… they skipped Blender over page formatting? WTF? (I realize this is probably not the case, but they sure make it sound that way…)

Not that I don’t agree with Ton that Blender GSoCs haven’t exactly gone great in the last couple years.

Honestly, I think it was a convenient excuse myself. As Ton said, it hasn’t changed over the years. That said, Blender’s page doesn’t really meet the requirements as stated on the GSoC page. No list of potential mentors, the projects don’t give “a list of prerequisites, description of programming skills needed and estimation of difficulty level”, and so on.

Combined with the lack of success, I think Google most likely came to a similar conclusion as we have - they weren’t getting value for money. Projects were not being finished, projects were not being merged, and the Blender effort towards the GSoC was looking a little like it was being taken for granted.

If you compare the number of accepted organizations within 2009-2015 (2009 - 150, 2010 - 153, 2011 - 175, 2012 - 180, 2013 - 177, 2014 - 190, 2015 - 137), you can see a slight decrease in this year. So, maybe they made a decision (?) about the maximum number of organizations/projects that were going to be accepted and Blender just found himself under the line (sorted e.g. according to quality assessment of ideas pages).

I agree with Ton. GSoC should be a privilege not a right.

EDIT: By the way, since there are organizations like Gnome, GitHub, Wikimedia, Ruby or Python on in the list of accepted organizations, the hypothesis about Blender being too big and well funded to be accepted is not very likely.

Perhaps one of the reasons, besides the ones mentioned in the above posts, is the fact that the current development of Blender not be consistent as a whole. We all know that some parts of this suite are privileged by BF / Devs and others (even if they have acceptance of USERS blenders) are relegated to ostracism.
And I find it strange a moderator / administrator to assert a user blender that he wants finaciar your pet feature blender, feel free. I thought this was a free software eque all opniniões were valid, not just a small niche of users.
Sorry if this is a opnion / very harsh criticism but it is how I see the situation around the discussions about the current desenvolveimento Blender. I accompany this fantastic software since the first versions and I feel (and other) frustrated with some attitudes of BF.
Anyway, long life to Blender!