Blender accepted for GSoC 2020

General Forums Blender Development Discussion
1

original email: https://lists.blender.org/pipermail/bf-committers/2020-February/050469.html

7 Likes
2

Wrong link to the mailing list

1 Like
3

Where can I find a list of Blender’s features that were developed/refined or grew out of the SoC programme?

4

https://wiki.blender.org/wiki/GSoC

Just click on link relative to one year to see a list of projects that happened, this year.

2 Likes
5

I hope that some student proposes projects related to texture painting, especially easy real-time painting with materials (with Eevee).

2 Likes
6

Fixed, thanks!

7

That’s planned and Pablo is going to take care of it because it will be a big project, I would rather see them tackle some modeling and uving parts, there is already code for those that just needs update and polish but i don’t know if GSoC/BF accept to use it as projects proposals depending on the mentors availability.

1 Like
8

I saw the “ideas” list, and I was like “wtf?!”.
Whoever wrote this list clearly isn’t using Blender in a production environment, there are so many obvious subpar areas of Blender that need fixing and improvements before considering all these “nice to have” features.

1 Like
9

It’s more about what a student can achieve than what blender needs.

Also the whole GSOC idea is about the Students learning and being introduced to open source, the project actually being merged in the ends is a plus as far as Google see it anyway…

9 Likes
10

Some of Blender’s weaknesses can only be solved with extensive design changes and a high level of coding experience. A GSoC student will generally not have the time nor the knowledge needed to resolve those (subsurf comes to mind).

Leave the fundamental design and refactoring work to the core team, as such a project is far more likely to be finished if a paid contract is involved. In fact, the core team should look at putting a bit more priority into performance and fundamentals as Blender sees more volunteer developers for the fun stuff.

4 Likes
11

To any potential students: An area with a lot of low hanging fruit is the UV editor. It doesn’t even have basic align/spacing tools.

6 Likes
12

There’s so much low-hanging fruit there that it’s a mystery as to why no one has appeared to take it on themselves to write patches.

This type of thing is a major reason why the current push for more volunteers is very important, but I could easily see a GSoC write a suite of basic, but missing tools for UVmapping (if he feels ambitious, add in a sync operation for selection).

13

Most of things proposed in User Interface section are just about finishing original 2.8 design, they had in mind.
That is the most filled section but probably the ones with features, users care less about.
But fixing all that stuff, now, will help them to avoid 2.8 original ideas only completed in 2.9 or 3.0 series.

Replacing Blender Internal textures still used by modifiers/particles/brushes/Freestyle/texture nodes by Cycles/EEVEE textures is something that should have been done before 2.80 release, too.
But they did not have the time. It will restore consistency between shading and modifiers.
Nobody can deny that would be useful.
But for a student looking at so many areas will probably be challenging.

Automated Testing will improve bug detection on dev radar.
That is not luxury when you see how short release cycle produces a fast increasing technical debt.

At last Blender Conference, Compositor was mentioned by users as a priority. And bugtracker was confirming that older bugs were about compositor.
That was clearly a promise made to users to look at it. And such GSOC would clearly respond to your demand.

Many Light Sampling and Fast Import/Export were GSOC projects in 2018-2019.
For some reason, the work done did not end up in 2.8. So, lost of the work would be avoided if it was updated to current 2.8 status to be implemented.

And eventually, VR stuff, Audio stuff, Bevel modifier, Mantaflow stuff : that corresponds to domains of devs available for mentoring GSOCs.

That is not a very exciting list from user point of view. But At least, that is a realistic one if you want to merge quickly your project in official Blender.

Anyways, this list is just a list of suggestions.
Students can have different ideas and they can propose something else.
And generally, among accepted proposals, there are always some of them that are corresponding to community needs, exciting and not from list of suggestions.

14

Supper important that the task be something that can be completed as a feature. So setting a task to improve a feature is a good idea.

An example of something that should not be attempted is the Mocap tools. Still unfinished. What, over five years now?

It should be something in the roadmap. A small hanging fruit to add. Such as one aspect of UV mapping, mentioned.

15

We have had two(2) summers of code development for LANPR, yet it remains unfinished!! and the Key Developer is under some kind of Chinese Internet containment (boo!!) It needs a determined push to finish, and get the features and documentation for more artistic usage. Freestyle controls, integration with Grease Pencil, modifiers, etc. Speaking of Freestyle, which I enjoy, could it be made RealTime rather than Post Process?? Until its features are integrated into LANPR it could use a speed boost. Some better integration of NPR texturing. The BEER initiative languishes due to funding?? might be a useful target?

2 Likes
16

If it was simple, they would have tried it during 2.8 alpha period.
If there was a second GSOC on LANPR about conversion of strokes to GP objects, it was to be able to customize them through modifiers like Freestyle strokes.
If Freestyle could be made realtime, @xp8110 would have done it instead of developing LANPR.

17

Ideas. Make at least a research on what tools are missing for:
1.Modeling
2.UV Unwraping\Editing
3.Baking
4.Viewport etc.
These topics very rarely pop up in the plans of developers, but the game industry is not smaller than the movie industry.
At a minimum, do research on these issues. But this is a coding quest …

18

I know most of these can’t be considered for GSoC, but I see no mention of these in the developers plans, or I missed them. And to me it should more a priority rather than VR tools or else. Fix the foundations before adding more floors.

  • Normal editing tools with precise controls (numerical and manual). For video games it’s a must-have, even for people into NPR…
  • Proper UV/packing tools.
  • Steamlined and enhanced baking tools. It “works” but it’s bare bone and not always successful.
  • Vertex paint improvements (like Vertex color master addon, but we’ve layers). Was it confirmed Pablo is working on this?
  • Streamlined and efficient retopology tools
  • Better snapping features
  • Better texture painting features (layers, brush management, etc).
  • Universal addon library/manager/repository, addon creators would share their creations through it, better visibility and access, checking for updates automatically, etc. Or something like the Unity/Unreal asset store directly inside Blender.
2 Likes
19

Thank you for your thoughtful response.

20

Well I don’t know where you looked, but much of what you ask for is evidently planned to be worked on, just not in the short term:

That leaves tree out of eight areas you mentioned, for which nothing is planned afaics. Also, as you basically mentioned yourself, hardly any of the above is simple and self-contained enough to qualify as possible gsoc project. Much of it leans into the area of design-tasks etc., even (e.g. better baking-workflow).

greetings, Kologe