I have no idea what you mean by the too many captains. GSOC is mostly about giving students the opportunity to contribute to an open source project and being paid a little bit for it.
The people who mentor the students are usually from the corresponding modules. If the code quality is good enough, they could review it themselves and introduce it to master very easily.
I wonder why that doesn’t always happen?
Incomplete work with implementation with the rest of Blender, not just the module itself. After GSOC for Texture Paint, it still took quite a bit for Psy-Fi to finish the implementation and re-evaluate how some things should be done in that implementation - the GSOC was the initial working mockup, and the addition to master seemed to be all the problem solving and workflow adjustment after.
Sorry if I wasn’t clear, that was a rhetorical question at the end.
It is just naive to assume that GSOC projects get merged at the end. Of course, it has to be the goal and the Blender developers have learned a lot over time about how to increase the chances, like by only allowing relatively simple projects to participate. Despite that, it still takes a lot to contribute good enough code. And even if the project is finished according to the initial requirements, there may be parts of it that are simply not good enough.
guys, I apologize for being distant from the Blender dev community, I’ve been busy with school exams and other things happening in my life. I didn’t compile blender for a year now, and so I’m outdated. Anyway I wish I could have a lot of free time to go back to this project and merge it to master. The problem is that if I aim to deliver it to master, I’d have to be active to maintain it for way longer, which I’m not available currently.
If anyone is willing to finish and polish what have already been done, I documented quite a lot what I did, and I’m available for explanation.
Here are some resources to understand what was the process I used :
Maybe, and I got burned by that thought with the Outliner project a couple of years ago.
LOTS of people contributed to the conversation on the dev forum with great ideas and well needed enhancements to the outliner.
Only to see some half baked options to show up in the next releases.
There seems to be little to no love for user suggestions and critique at the BF HQ. I see it often at the dev forum, as they’re just scoffed with a silly answer.
But that’s me, old man shouting at clouds.
There is no shortage of good ideas. Throwing more onto the pile doesn’t help that much.
If you wanted to take that idea, turn it into code that meets the standard of quality needed to ensure that it doesn’t bog down the rest of the code base, and submit that as a patch, it’ll be much more valuable than just an idea.
True, but having GSOC’s and then do nothing with it just silly.
Maybe, just maybe someone at HQ could look at the top requests and see how much work it would take to improve/implement it.
A while ago they turned down the speed of new features and focused on bug hunting and fixing.
It would be interesting to do this for most requested features.
Make a point release with just that?
Why should GSOC be treated differently compared to any other contribution? If it is not done or not yet good enough, why should it get a special treatment? This makes no sense to me.
I guess your complaint boils down to “The developers are not listening!!!” and it is less about GSOC, except that they didn’t finish a project you liked.
If you find a reliable methodology to find you which features are most requested by the community, chances are high you would get a Noble price for that!
Critic on Blender, the devs, development or the way it is all managed in any form is still a no-no here it seems.
I gave my opinion, and I’m out. That’s all folks!
I am a huge fan of constructive criticism. Or also if issues are pointed out that were not known so far, I am fine with that too.
“The developers are not listening”, without any suggestion as to how this could be achieved in the real world is getting old to me.
Dr. Sybren remarked on the development forum that the purpose of GSOC is to give students an opportunity to learn to code, contributing to an open source project.
Part of the learning process may ultimately result as “this is not how it should be done.”
If something from that makes it into final, that’s great. It would be awesome if more did. But I think it helps to temper the expectations if we remember what the point of it is, which is not simply free labor to create release-level improvements.
ETA: though I will add that I believe it would be ultimately more beneficial to the user base, if some of these projects were not bikeshedded to death.
Hosting a GSOC student on your project is essentially community service. It helps the open source community as a whole way more than it helps the individual project.
If you haven’t yet set your expectations of any gsoc project getting integrated into blender to zero, do so now. It will make your life much better.
This is a selection of changes that happened over the last week. For a full overview including fixes, code only changes and more visit projects.blender.org .
Anim
Make creating slotted-action pose assets assign the slot (commit ) - (Nathan Vegdahl)
Make pose library pose blending work with slotted actions (commit ) - (Nathan Vegdahl)
Support ‘Stash to NLA’ on slotted Actions (commit) - (Sybren A. Stüvel)
Disable ‘New Slot’ operator on linked Actions (commit) - (Sybren A. Stüvel)
Compositor
Implement single value reduction for CPU (commit ) - (Omar Emara)
Implement previews for new CPU compositor (commit ) - (Omar Emara)
Implement CPU domain realization (commit ) - (Omar Emara)
Inline 2D threaded for utility (commit ) - (Omar Emara)
Add stub CPU implementation for nodes (commit) - (Omar Emara)
Core
Replace home environment variable access with BLI_dir_home (commit ) - (Campbell Barton)
Curves
Support View3DOverlay.display_handle property (commit ) - (Laurynas Duburas)
Docs
Simplify the description of the select by pole count operator (commit ) - (Campbell Barton)
EEVEE
Improve GGX NDF precision at small roughness (commit ) - (Weizhen Huang)
Update the GGX LUTs (commit ) - (Clément Foucault)
GHOST/Unix
Always use getpwuid for home directory access (commit ) - (Campbell Barton)
GHOST/Wayland
Reference the stable tablet-v2 API (commit ) - (Campbell Barton)
This is a selection of changes that happened over the last week. For a full overview including fixes, code only changes and more visit projects.blender.org .