Like moths to the flame… 
Can’t stay away either, only to go “oh boy…” after reading on for 2 minutes.
1 Like
What exactly do you mean by stability?
1 Like
Well, unless I missunderstand what you mean by ‘create custom modifiers’, allowing such a thing would basically imply supporting something like a c++ api.
Why that was deliberately decided against, can be learned here:
https://wiki.blender.org/wiki/Reference/AntiFeatures
Regards your second point, I see @thinsoldier answered that in the meantime. His remark with respect to your first complaint is on point either (along with what I expressed above).
You can ask me politely not to. But it is not for you to command me to do or not do whatever you prefer.
greetings, Kologe
1 Like
Unfortunately this thread attracts ignorance and self-entitlement like shit attracts flies. We’ll see how long before everything gets out of hand.
greetings, Kologe
4 Likes
The title is already setting expectations I think. More of a venting thread than a constructive informed proposals, but that’s OK I guess.
Not sure how others are doing it, for me it’s usually “Please don’t engage…” → “OK, maybe I just write this one thing” → * Immidiately regrets doing so
2 Likes
To be fair, fanboy culture surrounding Blender is still around, the difference is that it is now often found outside of BA (Youtube comment sections, ect…).
This forum does a good job of keeping a lid on it, but this is not the only place for Blender users to gather for better or for worse.
4 Likes
I’m not sure there are a lot of blender fanboy anymore, at least I don’t see a lot of them on BA nowadays.
I don’t see blender as a perfect software, and in fact I find some questions here interesting.
Having a plugin system is an interesting idea for instance, and why it’s not possible / done is even more interesting I suppose.
What really bothers me is the lack of respect towards developers, like blender devs own us something because we use it. And that makes me even more angry especially since many of developers spend a great deals of their free time adding new features to blender so everyone can enjoy them.
It’s because of such contributions that we have grease pencil, dynamic topology, Bmesh (Ngons) , the (outdated) particles system, many recent sculpt addition, animation nodes, many new geometry nodes, and so much stuff that it’s impossible to keep track of everything.
“I find insane, that developers didn’t fix that feature yet on their free time , so I can enjoy my free time doing poor artworks ! Especially since I stopped using XXX commercial software, which was obviously better …”.
If that was non-blender users I could have understand, but if someone spend some time writing here it’s that they use blender regularly, in that case why being rude with people making it. And why not admit that we are all part of the same community, the development is open, we can follow it.
Why can’t we all do a little something constructive here rather than just whining ?
Is that what you call fanboyism ?
14 Likes
First, I’ll answer the second question, why not from the start? Because Blender was not architected to be extensible by default. This is a pain point for development in many areas and over the past few years a lot has happen to make Blender more extensible/object-oriented.
Second, why no custom modifiers? Because Python is used for the API, and core Blender developpers fear that too many custom modifiers in a file may result in slow depsgraph updates. This is because Python requires that only one thread at a time can executed Python code, so if, say, two objects have custom modifiers, they could not be evaluated in parallel. To me, this is a bit hypocrital as the fluid modifier is de facto a Python modifier (Mantaflow is accessed via Python).
I feel like custom modifiers should be a thing, and I do have a patch for it (if anyone is interested), but it is unlikely that it will ever be included in Blender because of Python. Further, if custom modifiers cause slow downs, there is still the option to bake to Alembic/USD/etc.
We could just “expose” the mechanism used for Cycles. It would be the same API as for Python, but in C/C++. As a matter of fact, I also have a patch for that, and started making add-ons in my custom build using my own programming language. This is also not something that may be accepted by core developpers.
5 Likes
Ummm… that sounds a bit contradictory to me. Geometry Nodes ARE custom modifiers, that is their literal purpose. The long term goal for Geo Nodes is to replace modifiers: https://projects.blender.org/blender/blender/issues/86839
Now I think a legitimate point for insufferable bitching would be that it’s taking quite some time. Not even the most popular modifiers like solidify, bevel or shrinkwrap have been ported. They can’t even “talk” to each other, for example: geo nodes can’t write a bevel weight to an edge for the bevel modifier to work on it, and the modifier can’t read edge attributes written by Geo Nodes and apply a bevel to them.
9 Likes
They are all being worked on either by the devs or the community or both.
The “even thickness” feature of solidify is very very tricky. 2 Devs are working on a more intelligent bevel algo as a Geo Node. I think I’ve seen node groups in the wild that are like some of the shrinkwrap features but not all. In some ways better, in some ways worse. In most cases just tailored specifically to what the artist was working on at the moment.
2 Likes
Seems the idea is to turn Blender into a Houdini and it will be a failure.
You cannot have wide appeal with Houdini very narrow band and high threshold way of working.
1 Like
I suppose that once everything is extensible by nodes, a new GUI will follow that is able to hide the complexity of the node system behind simplified application (like taking a bunch of nodes and box them into a group with a defined access panel and ports, but on a higher GUI level). So, you would not need to go through 40 nodes for an effect, you just take the box out of a library and apply it. Which would be similar to applying a legacy modifier. If you need “more”, then you can open the box and play with the content, or do your own boxes.
The critical part would be to find a balance between an accessible, fast interface with lots of drag and drop and automatic functions, and the option to go deep into the system and modify (and debug!) the inner workings. Both extremes have their challenges, and the combination even more so.
6 Likes
Geonodes have been a success for two years already, so that prediction is moot. The plan as far as I know is to turn every legacy modifier into a geonodes equivalent, and ship it with Blender.
2 Likes
How that success is being measured?
With scanning electron microscopy
9 Likes
So it is just your opinion…
Python is a big hurdle for performance, the future of Blender might be a fork. 
A number of very useful nodes (like Attribute Smooth) are contributions from the community. The people concerned about the state of nodes need to take into account that devs. like Hans are not just working on new nodes, but on a fundamental rework on how Blender itself operates on a core level. People have probably noticed that Blender is a little more unstable than it usually is in the last year or so and that work is the reason why.
When everything is done, Blender should be faster and far more powerful than before, but like in every other area, it is not as easy as just throwing more people at the problem.
5 Likes
In my opinion, the hair assets are a great step in that direction.
2 Likes
It is great to see they are taking the opportunity to rethink and rework many details, instead of just trying to finish it as fast as possible.