Module Teams

The Blender development team keeps growing, both with Blender Foundation developers and external contributors. To scale the project we need to organize ourselves better, by avoiding complicated dependencies and bottlenecks between projects that we had in 2.80.

At the Homestretch workshop last week we talked about this subject and decided on some steps forward.

The first part of the solution is sticking to a fixed 3 month release schedule. This means each project then be developed at its own pace, and go into an official release quickly once it’s ready. Every feature must be stable when it goes into master. If not it moves to the next release and does not delay other projects. We intend to enforce this strictly.

The second part is establishing active module teams that take responsibility for their respective areas. Previously module ownership was divided in many specific parts of the code, in some cases with no clear point of contact or communication on the overall direction. We have now defined 11 modules, with experienced developers responsible for coordinating approximately one module.

Modules on wiki.blender.org

This page will be updated in the coming months to provide a clear overview of what we are working on or planning to work on in each module for 2.81 and later. There should also be a list of tasks that new or existing contributors can help with, similar to the successful user interface paper cuts project. Features will not be attached to specific Blender versions until they are stable, rather they will be marked with priorities.

The Blender project administrators remain the same. I will have the role of reviewing designs and architecture for the Blender project as a whole, to ensure there is a consistent direction, decisions get made, and that projects have a well-defined scope that fits in the new release schedule.

19 comments 6,113 Views
  1. As the Development fund reached 30k, can we expect new developers to join the team?

  2. Oh I totally missed Dev fund getting to 30k.. Amazing!

    As for modules sounds good. Hypothetically means that everything from mantaflow to Udims to sculpt branch features could be in master 3 months after 2.8 release if miracles happened. In either case sounds like it will helpfully help with getting new features to master faster.

  3. ” If the development fund grows beyond 30k:

    Particles and hair nodes
    Texturing tools and tools for procedural textures
    Painting and Sculpting improvements
    Better snapping and precision modeling
    Cycles: denoising
    Compositor ”

    Can’t wait…

  4. man, if i had a million dollar spare, i would have donate it asap to blender’s developement……but like everyone else i am just a little nobody…we all should contribute to the fund even if it’s just a small fraction, keep up the good work guys.

  5. Blender has already 6 full time paid developers so there is your 30.000 euros
    If we reach the remaining 20.000 euros we will get another 4 devs

  6. Note Jeroen Bakker was hired recently which makes it 7 full-time developers, and Sebastian Parborg a bit before that to make it 3 other developers part-time. But yes, to hire a bunch more developers the fund needs to grow.

  7. Sounds good. This will undoubtedly improve Blender development.

  8. Holla Brecht,
    Most of the best Blender features are brought into Blender with a good collaboration between devs and artists/users, it will be really good to have an artists in every module. To make sure that work-flows don’t break.

  9. i like this module teams concept but who decides on the new features, the module owners , the whole team, the Admins..etc? who makes the prioritey of which features are important or not?. i know u guys take community feedback/requests into account when it makes sense, but who are these “User Members” that are appointed by invitation by the module owners….these points need clarification so we can better understand how the roles of each individual plays…thank you.

  10. while i admire what you guys are trying to do with this version, but i feel we missed a great opportunity in terms of improving performance, handling dense meshes both in Edit/UV and Sculpt modes (for example 3ds max can handle millions of polys in those modes with no performance hit and the undo system is very responsive)”autodesk is doing the same for maya”, real construction history, realtime playback with complex rigs and many more, these are the core and important issues that should have been tackled in every major release, i understand that the UI/UX needed redesign which we all agree , but we didn’t need EEVEE at least not right now ,until these big issues are handled first, trust me when i say you’ll see more problems, crashes, user frustration…etc this engine will cause more problem on people’s machines even with a small scene…..also 3d softwares companies are making their viewports closer to their off-line renderers and preparing for a fast previewing and rendering instead of developing a game-engine like that relies on approximation which is not even at the same power of a AAA game engine like UE4 that most of them have a direct live link with…..i am afraid u guys are heading in the wrong direction and merly focusing on what’s ‘flashy and cool”………i hope you guys take my comment seriously and not get offended and see what can be done to reshift the focus to these problems…thank you for you attention.

    • I fully agree with the performance issues being important, and I realize it must be even more frustrating for long time Blender users who were perfectly happy with the user interface and usability.
      On the other hand UI and usability changes are what’ll bring Blender closer to a broad audience. EEVEE was great for marketing purposes as well as Grease Pencil. That made the development fund go up significantly which means more developers who can focus on these important issues.
      Bringing performance up to par with 3dsmax or modo on it’s own wouldn’t do much to win more people over and get more funding.
      I’m sure that after the 2.8 release they’ll focus on those subjects as well, I’d be as happy as the next guy to get a fast edit mode performance.

      • lets hope we don’t have to wait for another 5-10 years to get these fundamental issues tackled, the way i see it, is like building a house, if you don’t set it up on a good foundation and jump straight to painting and decoration, everything will evenutally crumble, and not only the people you tried to bring in will leave but also those who stuck with you for a long time… i am optimistic because i believe if they want to do it, they can do it…..it’s just we don’t have to wait for years and years ,version after version to see it done.

        • Absolutely. Now that they’re more in the center of attention it’d be a good time to address some of these issues (easy to say, I know). Especially since professionals/studios can’t afford to wait anyhow, it either works or it doesn’t, mouths to feed, bills to pay, etc.

          They seem to be aware, interested to see how this ticket turns out – https://developer.blender.org/T57936

          But even if it’s bringing back edit mode performance to 2.79 standards, that might not be enough. Not sure how complex the issue is, haven’t tried 3dsmax in a decade and modo in 5 years. Blender 2.79 is faster than maya for high poly modeling/deformers so that’s already a selling point for me – that’s not saying that editing anything past 100k polys isn’t a pain.

      • “Bringing performance up to par with 3dsmax or modo on it’s own wouldn’t do much to win more people over and get more funding.”

        It wouldn’t chase people away either. For sculpting and texture painting it’s almost mandatory to use Zbrush and Substance/quixel/mari if you want to do anything moderately heavy and that’s where people are going to. People who don’t use them either stick with stylization or they have some immense patience and tolerance for inefficiency.

        • I was saying that in the context of Eevee and Grease Pencil being good features for marketing purposes and expanding the user base, since it was mentioned before that performance regressions are a more pressing issue than an improved realtime viewport. Obviously noone has anything against improving things.

          I agree about Zbrush and Substance not being replaced anytime soon. Their moderate pricing for perpetual licenses makes them also convenient for freelancers to use – but Blender might transition into that territory. As it is now I’m using substance for texturing, but for smaller tweaks, touchups and let’s say an additional quick map for sss I’m suing Blender. While I use zbrush for sculpting, smaller sculpt tweaks, getting rid of intersections, tweaking poses, proportions can be done in blender more easily instead of exporting/importing, even with goB.

          I’m sure it’s a matter of personal preference, but with an improving feature set I can see myself doing more and more in Blender, things I would have done in other packages before.

    • i think you got the order wrong.

      you can’t optimize performance and then do big changes like EEVEE*. you first have to do EEVEE* and then you can optimize performance.

      (* and the other things blender 2.8 will bring. like new the dependency graph and so on…)

      don’t be worried! 🙂 the blender developers know pretty well what they are doing.

      • i am talking about internal things and how blender does sutff at a core level which u can only break when there is this big upgrade that happens once in a decade, for example i think that blender copies the whole scene in edit mode which causes huge bottleneck and slow undos or something…these things are unrelated to optimization or EEVEE but rather at the core systems which should be also changed even if they break compatibiilty, otherwise you’ll never go beyond what’s in old versions and that’s why other softwares have the edge over blender in those areas.

  11. 1)Materials and textures.
    2)Materials and textures.
    and
    3)Materials and textures.

    Thank you for your work.

  12. Hello and Good afternoon.

    I wanted to ask if there might be any plans or ways to easily produce scaled documented drawings directly in blender.
    In themed design and other fields it would be great to easily create front, top, side views, section cuts and other annotations from inside one program.
    It would be great if there was a way to easily and quickly generate build drawings from a sculpted model inside Blender would change my life. If there is anything I can so please let me know to help develop this topic.

    Kindly

    Travis G.

  1. Leave a Reply

    Your email address will not be published. Required fields are marked *

     

share this story on