To recap your argumentation, it would seem to me:
As an argument in favour of what you want, you consider sufficient:
While as a counter-argument against what you want, you’ll only accept:
greetings, Kologe
Get rid of hard-coded/custom attribute in favor of a generic interface.
You might have noticed that you could have only 8 different vertex colors and only 8 UV layers. But a very high amount of Vertex Groups. You’ll see that color attributes don’t have that limitation.
If all is just generic float, 2D , 3D , color, attributes, they all fall under the same system with same optimization and limitations. Making them much easier to maintain therefore making the overall system more stable.
Lol, I think you’re missing the point. Vertex colors worked for sure, but the way they were implemented was getting in the way of better overall performances and better architecture design.
It’s not about making VCol better, even if they’ll get instant benefit from amelioration to the attribute system, it’s more about making the whole mesh and attributes ( Vcol, UV, Vgroup, FaceSet, Smooth, EdgeSharp…) much efficient.
There as been a lot of complains about poor mesh performances, and these kind of global refactor is a step in addressing these issues too.
If you want in-depth rationales about changes in the code, you can follow development tasks : https://projects.blender.org/blender/blender/issues/100153
If you want a discuss: please read all messages, not only a half of them.
I already clearly explain myself: changing names and ideology while we STILL didnt have ability to deal with vertex color alpha without addons mean what we dont fix or upgrade something, we just change the naming.
If you undercook the pie - put it back to oven and make it right. If you think what changing the title from “Apple pie” to “Apple attribute pie” will change something - i have no idea how to explain you where exactly you are wrong.
No. Those guy telling about some instability and optimization. That was his point:
benefits seems pretty obvious : less bugs, more optimizations
So where exactly those problems with optimization and bugs with vertex colors?
Not might. Exactly - break.
You just need to use a brush with an Erase Alpha or Add Alpha blend mode.
By default, transparency corresponds to a Black color in solid view.
You need to set-up material and use material view to work with transparency.
When an addon is obsolete, it often does not work anymore with recent release.
It is a way to communicate that they are attributes, that can be tweaked, generated by geometry nodes.
And that other types of attributes can be converted into color attributes.
That is also a way to inform people that things have changed like sozap explained.
During development, giving different names was a way to distinguish old vertex color system from new one.
what your addon do exacly ? can you share a link to it ?
Color and alpha is set by a color attribute using vertex paint :
I think you’re upset and confused because you don’t understand the changes and how development works.
Different things append in parallel, your issue is most likely in the painting / brush department.
Changes are planned in this area too, it’s just not happening at the same time.
While the change to color attribute is more about how mesh data is stored, how we interact with these data is another problem.
It shows you your alpha channel in vertex color. https://github.com/andyp123/blender_vertex_color_master
Dont tell me i can see my alphas by adding some materials manually.
You just need to add a Color Attribute node to material and then, plug it into Material Output or Principled Shader, according to what you want to inspect.
If you want color and transparency visible while you paint only one color attribute, you can plug its color and alpha outputs into color and alpha input of Principled Shader, set Blend Mode in settings panel of material tab to something different than Opaque, and enable Material Preview display mode instead of Solid one.
Then, if you want to inspect each channel individually, you use a Separate Color node.
If you want to paint only a black&white channels, you can create a material using multiple BW color attributes, plug them into a Combine Color node, and bake result into another Color Attribute channel to keep on exported mesh.
Misc operations can be assured by color filters and brushes.
The whole addon abilities are present in default blender, without using addon.
Nah, I’m stuck in a vfx pipeline, so for us it’s gonna be 2020 for another 5 years I guess. 
Internal tools will make anything unstable, so at least there’s that for comfort.
Well ! What about making a geo node setup that emulates that functionality ? It looks possible to do something there.
Or , what about updating that addon ?
It’s probably a easy fix and if you don’t know python you could have find help on the forum,
But probably a bit less with that whining attitude.
This is so-called API stability.
Blender doesn’t have this.
For a big production, this is a really serious issue.
Couple of months ago I saw this video on YouTube where some old/middle-aged man showcased some sort of CAD modelling tools in Blender, which was supposed to work better than Blenders internal tools, which were alien and an insult to humanity. He showed how he could do some object in 5 minutes.
I was watching interested about speeding up my workflow and when he mentioned 5 minutes I went huh? and tried making it with internal tools did it in less than 1.
Couple of months ago I got interested in Pablo Dobarro drama and read Big Sculpt Mode 1, and reaction to all of his novelties, people were outraged about this new workflow changes, which would make Blender unsuable for sculpting and for this and that. And I saw this one sensible comment saying maybe future 3D artists will use Elastic Deform instead of Grab, Remesh instead of Dyntopo and etc.
And I, who started Blender 1 year ago were using exactly that workflow. Some things mentioned here seem so alien to me, I pray for god Blender doesn’t fix those “issue” you old folks moan about. What the fuck has this thread come to. I almost don’t want to engage anymore the way moderators allow absolute filth and disrespect in here.
Today I’m working in studio, using tools you lot would deem abomination 2 years ago, make a living and having a blast of the time. I moan here a lot, especially about painting, but my god grow up.
Most of the arguments here goes like this:
Grow up.
This isn’t complaining about Blenders UI or bat-guano-insane things, all this (and Big Sculpt Thread mostly) comes to is complaining that Blender doesn’t validate YOUR workflow, and instead goes in other way. As a Blender newbie, again, I pray to God most of changes proposed here aren’t even entertained in developers head. I started this workflow, I learned it, find super comfortable, same way you guys did back in the days.
Some people don’t like changes, prefer old ways, no big deal. Blender allows you to change workflow, incorporate addons, even literally rewrite software to fit your requirements. Forum is filled with helpful people who’ll help you do that. GitHub and this forum is FILLED with absolutely free addons that straight up make Blender completely new experience. You can choose to do so and enjoy your workflow, but most of you don’t, because you don’t want fixes, you want Blender devs to validate your workflow, your knowledge, years of money wasted on paid softwares to publicly announce that your workflow is superior and will be default.
Its not.
And people who have this refusal to download .zip files and enhance software… my god, just grow up.
API is evolving, following growth of software.
That is perfectly normal and is happening with all type of software.
We are living a special period in cycle of development.
Since 2.80, developers are refactoring Blender fundamentals.
2.8x series was a first pass of refactoring with new features. (EEVEE, Grease Pencil Object, new Sculpt Mode, Library Overrides,…)
2.9x was like a pause, a pass of polishing.
3.x series like a second of pass of refactoring with new features (Cycles X, Asset Browser, Attributes, Viewport Compositor,…).
4.x series should be about simulation nodes, new Paint Mode, rigging nodes… There will be fundamental changes, but that should not be about basics of mesh structure.
Changing vertex colors into color attributes is something that does not happen, every 3/4 months, at each new release.
That happened once, and that is supposed to stay as is, for years, a decade, maybe more.
If you stick to LTS, you have 2 years of API stability guaranteed.
Schedule of LTS releases is public.
A big production will hire artists, knowing how to use current LTS, and a technical director in charge of maintenance of pipeline.
Such guy, knowing about LTS releases cycle, can easily anticipate an update of intern tools using API, by reading release logs.
What has been a problem in past was the change of python lib version, more than API.
It took years, but developers agreed on following VFX platform specs.
So, no. That is not a serious issue for a big production.
That is a problem for small structures with less resources, trying to adapt their pipeline to fast changes, happening every 3/4 months.
It is not normal.
Do you think a studio will rewrite their tools every 2 year?
Blender dev can add new features without breaking the existing API.
Yes, making API stable is not free.
But, Blender will never be a back bone of a meaningful size of studio without it.
No, Blender devs can not provide an API consistent with a refactor, without adapting this API to refactor.
They adapted the API because the name in the core had been changed.
Before 3.2, Blender was supporting only Face Corner Vertex Color.
After 3.2, Blender is supporting two types of vertex colors (Face Corner or Vertices). The system is becoming more complex.
Some part of software, that was specific to Face Corner Vertex Color, stayed specific to this type of vertex color. Another part of software, that was specific to Face Corner Vertex Color, became more generic, able to handle both types of vertex colors.
They had to modify API because the software was modified.
They did not change the API just to annoy addon writers.
They chose to rename the feature Color Attributes, because, in future, it is possible that Blender could handle Edge Colors or Face Colors, too.
You can not provide an API giving access to new features, if you keep old API ignoring those features.
They are free to do what they want.
They can choose to keep an LTS release of Blender during 10 years, if they want.
They are warned that Blender Institute will only provide support for this LTS during 2 years.
They are informed that Blender is in an expansion phase of its features.
Nobody is forcing them to use Blender, if conditions announced are not suiting their needs.
But if they choose to do it, they are probably prepared (not to rewrite entirely all their tools) but to update some of their scripts, regularly. And, or, to work with several releases of Blender in parallel.
It looks like you don’t even know what the API stability means.
There’s a bit of negative communication going on here- which is understandable, as this is a thread for venting; however, please keep your frustrations focused on your experience with Blender and not on other people. We’re all on the same team here; Team Blender, so please be constructive as you’re discussing your opinions with other people on this thread. Thanks!
What are you referring to exactly ?
Are you working on a big production and had problem with blender because of API changes ?
Can you talk more about it ?
To me, most of the time the changes are really small and you can update all the tools really quickly.
As @zeauro said , studio might prefer to stick with LTS if they prefer stability. But I must admit I
generally don’t work on big productions.
As a whole, it’s a chicken and egg problem, I agree stability is important, and who like when stuff break anyway ?
But on the other hand, a lot of things is needed in blender to make it useful to a “meaningful size of studio”, if that means something.
In that regard big refactors are expected and it will probably settle at some point…
Maybe blender isn’t for big sized studio yet, in fact who cares ?
Do you think a lot of people here are from big studios waiting for blender to be good enough ? As Ton said, if they are really interested they can help by contributing to blender…
But that’s another debate…
My question is, can you share some experiences about when these breakages caused issues and what happened exactly ? I personally have one thing in mind but I’m curious about your feedback !
Simply, you don’t change unless you absolutely have to.
If you change, you provide a way to backward compatible.
The fact that I have to explain this proves Blender is not for big production.
Well, sure, I can understand that to some extent.
It’s not the only argument against blender integration in bigger pipeline.
Even if you’d be surprised to find it in a lot of companies even big ones, but indeed not as their core tools.
How should I interpret the fact that you didn’t answered any of my questions ? Are you working for a big studio ? doing what exactly ? and how API changes in blender caused you some issues, how did you solved them ?
