Developer Meeting Notes

I don’t think that’s true, but I guess we’ll never find out.

(Same story with Rightclickselect, but cleaning that up is made even worse by the nightmare Ux of that site.)

1 Like

All it takes is someone to look at the most popular ones and make a decision. All cannot be done. Many should not be done. Some are no-brainers. Some are very good ideas but something never done before in any app and would require expert UI designers and developers to do it right and the entire world has been living without it all these decades so such ideas aren’t worth the investment.

1 Like

You are right that what can be done is to look again at the most popular ones and trash everything else, no matter how good the ideas are or how valuable they would be in practice. That’s very unfortunate and I am sure many are getting the impression they wasted a lot of time.
In my opinion, it was poorly managed rather quickly and quickly degraded over time. And I didn’t expect anything else, because someone would have to invest a huge amount of time to manage and structure it to preserve the value over time. To be honest, I am not even sure whether it would be worth the time to do that.

1 Like

A while back before the subscriptions, Adobe had a “just do it” initiative where they just started fixing a bunch of tiny annoyances in their main programs. I firmly believe that this generated a lot of positivity in their user base and made the switch to subscriptions much less painful.

The Blender 2.80 UI redesign was great but an extra little bit of polish on the UI will bring incredible pleasure and productivity to a lot of people.

Having the node editor auto-scroll when you drag a node is ORGASMIC. The improvements to splitting windows more intuitively is beyond orgasmic. Finally being able to ctrl+mousewheel with a macbook trackpad is a huge time saver for me. The ability to export Quick Favorites as a python script and run it on another machine to get the same favorites would make so many people so happy. Movng the view pie menu to a key everyone’s keyboard has instead of ~ would make so many laptop users all over the world so happy.

There are at least 20 no-debate-necessary improvements than can be done to the UI that would make so many people so happy it would dramatically boost the number of Blender download for a few weeks, maybe even increase individual donations slightly.

That’s great, but I don’t believe that kind of thread is required for that. The developers are aware of 99.9% of the annoyances and there is only occasionally something they haven’t seen.

I don’t claim no more polish is needed. My claim is that the thread should have been closed way earlier. Likely shortly after 2.80, but I am not too sure about that.
It made a lot of sense for 2.80 where a lot of aspects were changed and the whole concept became “new” for everyone. So, getting help from the community at that point made a lot of sense.

I agree that such improvements exist, but to fix those, no such thread is needed.

1 Like

I was thinking some time ago about the problem of ‘Papercuts’ thread on Devtalk.
There is a solution to this mess, but it would require couple dedicated people that can do systematic work for a long period of time.
Basically the entire thread should be fine combed, all the proposals listed, categorised and pre-prepared in a digestible form for a developer. That means sifting out all the garbage, merging doubles, etc.
Since the threadposts are numbered, the thread can be divided into sections, so the team, lets say of 4-5 people can start working from multiple points (kind like tunnels are bored IRL).
Dedicated task on Phabricator could be made for cooperation.


How many man-hours of developer time do you want to sink into sifting?

I assumed (perhaps incorrectly) that this was about my recent area maintenance changes, so I read this line out to my wife. She retorted instantly with the most savage comeback of “I’ll bet nobody has called you that before.” Ouch.


Zero. I meant that this could be done by a team of volounteers, aka - users.


The next time I’m ****** ** ****'s ***** from the back I’ll have her say “HARLEY” during her third ****** in honor of your amazing improvements to the Blender UI.


Sweet fucking jesus, I can’t believe my eyes

Thank you Harley ! have your wife try this patch, she will know the feeling like I just did lmao


Would need to try before & after for proper comparison. :wink:



Modules and Projects

Geometry Nodes:

New Features and Changes


  • Support a new Local Space (Owner Orientation) for targets. (commit ) (Alexander Gavrilov)
  • Implement ‘Remove Target Shear’ and more Mix options in the Copy Transform constraint (commit) (Alexander Gavrilov)

Mesh Editing

  • Performance Improvements
    • Skip edit-mesh duplication for most BMesh operators (commit ) (Campbell Barton)
    • Support multi-threading for bend transform (commit ) (Campbell Barton)
    • Support multi-threading for most transform modes (commit ) (Campbell Barton)

Geometry Nodes

  • Initial attribute list for meshes (commit ) (Jacques Lucke)
  • Curve primitive nodes (Johnny Matthews)
  • Add mesh input warning to curve to mesh node (commit ) (Hans Goudey)

Video Sequence Editor

Grease Pencil

  • New modifier to generate weights dynamically (commit ) (Antonio Vazquez)
  • Add weight factor to Offset randomize (commit) (Antonio Vazquez)
  • Line art modifier
    • Occlusion effectiveness support (commit ) (YimingWu)
    • Allow filtering intersection lines using mask numbers (commit ) (YimingWu)
    • Use separate sub panel for line types (commit ) (YimingWu)


  • Reduce shadow terminator artifacts (commit ) (Mikhail Matrosov)

User Interface

  • Add the ability to see all view layers in the outliner (commit ) (Dalai Felinto)
  • Hide collection tab when scene master collection is active (commit ) (Yiming Wu)
  • Do Not Resize Already-Open Temporary Windows (commit ) (Jesse Y)


  • Apply random selection precisely for curves, lattices & objects (commit ) (Piotr Makal)

Weekly Reports


The edit mesh performance improvements are getting out of hand. Haven’t tested with these patches yet but even before them the difference was stunning.


  • Jesse Yurkovich (deadpin) joins the “Platforms, Builds, Tests & Devices” module.


Modules & Projects

New Features and Changes

Mesh Editing

  • Remove redundant mesh-backups from edit mesh operator API (commit ) (Campbell Barton)
  • Tag the mesh instead of the object when editing with number buttons (commit ) (Campbell Barton)


  • Constant folding for improved performance and memory usage (commit ) (Manuel Castilla)
  • Full-frame implementations of more nodes (commit , commit , commit , commit ) (Manuel Castilla)

Geometry Nodes

  • Replace the per-node view toggle with new viewer node (commit ) (Jacques Lucke)
  • Curve Endpoints node to output the start and end points of curve splines (commit ) (Angus Stanton)
  • Add explicit Float to Integer conversion node (commit ) (Nikhil Shringarpurey)
  • Rename nodes for clarity between mesh and curve (commit ) (Hans Goudey)
  • Curve Primitive Nodes:
    • Curve Primitive Line (commit ) (Johnny Matthews)
    • Allow 3 points in curve star primitive (commit ) (Hans Goudey)
    • Use same shape as mesh line in curve line node (commit ) (Hans Goudey)


  • New evaluation mode option like the USD exporter (commit ) (Philipp Oeser)
    • Remove non-functional “Renderable Objects” only option (commit) (Philipp Oeser)

Exact Boolean

  • Various parallelizations and optimizations (commit ) (Erik Abrahamsson)
  • Fix performance regression due to exact triangulation (commit ) (Howard Trickey)

User Interface

  • Node Editor
    • Add a button to groups to change the type of inputs and outputs (commit ) (Lukas Tönne)
  • File Browser / Asset Browser
    • Select files and directories after renaming (commit ) (Julian Eisel)
    • Keep assets active after renaming, ensure they are scrolled into view (commit ) (Julian Eisel)
  • Various improvements, tweaks, and naming changes
    • Outliner: use ‘compacted’ row for bones as well (commit ) (Philipp Oeser)
    • Center the Status Bar Progress Text (commit ) (Vincent Blankfield)
    • Rename Scene’s embeded collections from “Master Collection” to “Scene Collection” (commit ) (Bastien Montagne)


  • Tests: performance testing framework (commit ) (Brecht Van Lommel)
  • VSE: Use snapping settings for scrubbing (commit ) (Richard Antalik)
  • Make Single User: support object data animation (commit ) (Philipp Oeser)

Weekly Reports


From nodes workshop blog:

Geometry Checkpoints

For this we introduce the concept of checkpoints, a place (node) in the node tree that can be frozen. The geometry created up to that point is baked and can be edited directly in edit mode. Any nodes leading to the checkpoint cannot be changed without invalidating the manual tweaks.

We finally gonna get something like edit poly modifier from 3ds max? :exploding_head:


To be honest, while this is certainly an improvement over what we got up until now, it’s still not a proper solution. It makes the puzzle you need to solve to close/open editors somewhat easier, but it’s still a puzzle to solve, compared to normal, common sense interfaces with drag and drop customizable, tabbed layout.


It does sound like that!

It will be close. But, edit poly doesn’t freeze mesh. It even works on an animated geometry.

freeze means that, when you edit the mesh it will stay like that, unless you change the stack position the modifier, edit the geometry before it, or remove it…