Blender's increasing development pace and the lurking danger of overcomplication

In theory this is undoubtedly the case, but in practice I see lots of commits being code-wise evaluated and approved relatively quickly if technically alright, but not carefully judged by their significance.

I guess a good example is the impressively rapid pace of Blender’s Sculpt Mode development. On one hand I welcome this, on the other hand a clutter of tools and functions is forming. The significance of a number of those functions is disputable, especially when placed against UI / UX clarity.

Then there’s the constant shifting of UI elements and keyboard shortcuts. It starts to become a bit tiresome to me. I’d love to see the development pace slow down a little, and focus on quality improvement and feature refinement.

1 Like

Personally i wouldn’t mind if they release each version in 6 months instead of 3 or 4, this gives everyone room…Developers can add few features with a specific goal but the most important part is Polish and bugs fixing.
I think users who tend to follow the daily builds won’t be affected by this change and those who like to work with a stable version will get a great release each 6 months.

1 Like
A side effect of the "Keep X workflow / setup but accommodate Y now too" arguments. That's options for you.

Not complaining (about having options). Thanks devs.

Much like the downside to slowing the dev pace will be rotting patches and then the resulting conspiracy theories / threads about said.

Oh how I miss those. :cry:

Our current ‘problem’ is preferable. IMO. :goggles: :rocket: :checkered_flag:

1 Like

Maybe that’s just because you use daily builds, I think for people that only stick to official release it could be different.
Of course , when these people will switch to 2.81 there will be lots of new/different stuff , but they’ll have to adapt only once in a while.

IIRC during the 2.49 > 2.5 phase there was a lot of changes from release to release , because all the UI is different there is a lot of trial and errors. We are still on the polishing phase of some areas I guess.


True that, I also thought of this. Maybe I’ve become a little too involved. :slightly_smiling_face:

There has been lots of talk from time to time, if Blender should introduce a new plugin-based-architecture (or microkernel architecture) that works internally in the C codebase.

As for example nobody ever mentioned that “There are too many Python addons, people should stop making too much addons because they would complicate Blender”. But the real power and principle of addons is that you will dynamically load whatever you really need. So in that sense you have lots of control on how you can use the software based on your personal needs. Virtually this way you can load an unlimited amount of addons (or as long as the RAM can handle) but practically you will only load exactly what you want to use.

In this mindset, such an architecture on the source level, it would be like writing Addons directly on the C source level rather on the Python.

Currently, there is no such degree of flexibility, since Blender is built with a classic C architecture which is modular by nature, the only way to include-or-exclude features in compilation is only using control variables (preprocessor directives), but once something is compiled stays in the build.

1 Like

I think at a certain point the development (if speed keeps up) will sway again towards investing more time in code quality. A lot of new code also introduce new bugs / instabilities and unexpected behaviour…so I think at a certain point they simply will extend bcon 3-5 to a lonmger period of time. Wich will solve a lot of problems.

this of course will hurt people who do have the opposite symptom of feature fatigue: feature addiction.

As for overcomplicate… I think this is the fun part of development, encapsulating a feature in a well done workflow, wich on one hand is accessibole, but on the other can be extended to full potential by the need of the user. But I also think that this can be archived by a longer bcon 3 to 5 time.

But currently I’m all in, a new UI offers new possebillities, and current dev speed should be exploited as much as possible.

So I think this will sort itself out, if the prioties of the 2.7 dev cycles are still in place

my worthless 2 cents on this

1 Like

It seems you’re experiencing feature consumption. I can see your point. P.D. sculpt thread is moving so fast it’s impossible to follow. That guy speed is impressive…

What BF is doing now, it’s more or less, experimenting. The startup boost craze will probably go on for some time, until it will reach a peak. Then they’ll take a break.

I must admit I’ve skip 2.80 almost entirely. :stuck_out_tongue_closed_eyes: Just lately I’ve downloaded 2.81 for trying some sculpting. A lot of things I’m still on 2.79 :wink: it’s faster.

Take a break from the mess. Relax. You don’t need all that features. Use what you need. Blender has 3 different ways of doing everything. While looking one of those new sculpting features, I wondered, “Yes but… I can do that with deform modifier and a texture” :laughing: (joking… well no)

Thinking back, it has always been like that. I guess.

The most important things imho now is to get that Blender faster. More performant. If it has to break and hit. And that’s gonna be tough… being fast like a videogame engine while mantaining all that data…


If I remember correctly, some of the recent corporate contributions were towards project management areas like the ones mentioned. Good project management is the difference between “large ball of duct tape” and “functional commercial software” with large complex programs. It’s good to see those funds being put to good purpose. :+1:


What I’m concerned when it comes to the increased development speed and the big funding. Is that it feels like features and workflow and UI suggestions are mainly oriented towards big studios who are paying for them(makes sense). But in the end the little freelance user will have to swallow up what big studios have decided to be the workflow. Yeah there is rightclickselect , but I’m left with the feeling that developers seldom take ideas from that place.At least not in the past year or so.
But that might be because I’m a recent(2 years a go) blender convert.

1 Like

Honestly, I just hope blender polishes up those sculpt tools, gets a good remesher and maybe adds a little ergonomics to the texturing process.

1 Like

I say bring it on! The UI and performance guys may even catch up eventually. Go team Open Source! There is a reason why the motor industry throws so much money at the F1 Grand Prix, forged in the crucible that is Motorsport, so to speak…

I get that it could be problematic for those who always want to test-drive new features, but I don’t see any upside to slowing the rate down by intentionally adding layers of bureaucracy.

One reason is that the ability for it to slow things down could be too effective, making the BF like a overgrown national government who needs millions of dollars and a year of time to do trivial things like put a fence up (no joke, this was an actual thing with the US government). Instead of higher quality development, you would get a snail’s pace of features that take forever to get in, and once they do they could have problems that require a rewrite.

At the moment I’m trying to simply stick with the official release of 2.8 and look at daily builds at nice to tinker with. I’d rather see a update every 6 months with new features and maybe major bug fixes every 2 or 3 or when they are needed. I’d rather wait for a solid build than an semi tested feature. Feature creep can become a problem for software.
I’ve also noticed a bit of a problem with too many add ons and Blender shortcuts competing for the same amount of keys. Some add developers hardcode their keyboard shortcuts and they step on existing shortcuts, or ones set up by other add ons. I know you can set up work spaces and filter addons. But that just adds to the complexity. I almost wish I could strip blender down to nothing and simply add what I need. I’m sure one can I’m just not into coding. I’d like to see some sort of guidelines and some UI room set aside for addon developers so they are not all over the place like they are now. At this point making pie menus is the only way I can keep things organized.

Perhaps I would toss this out there as an (arguably dumb) theory as to why a sudden increase in Blender’s popularity and development rate is causing anxiety.

As Open Source lovers, we are very much not adjusted to an environment where our favorite apps. are coming to a dominant position over the majority of commercial options. Rathe,r we are used to having to defend the idea that our app. is relevant, used to seeing nice looking projects in limbo or outright canceled, used to losing and disappointment from time to time in general.

It became a comfort zone which the amazing new features, the massive new funding, and the wins in general is upsetting. We became so used to dreaming about the underdog going somewhere one day that we don’t know how to handle its transition to overlord.

There are other FOSS 3D apps. one can use that are still very much trying to find a way to become relevant, if the underdog world ends up being what you want.

:slightly_smiling_face: Adding layers of bureaucracy is not what I mean, but when a project grows, you have to be cautious it won’t get out of hand and starts living a life of its own.

I’m not against new features, and I’m glad Blender development is well-oiled, it’s just that ‘development’ does not only mean adding and changing until you drop, but also taking the necessary time to make sure Blender keeps offering a stable, clear user experience, and limit feature additions and UI changes to those that really improve the UX. That’s the point I’m trying to make.

A second episode of Cosmos Laundromat was announced at BC.
That will be a period of evaluation of 2.8 monster.
If priorities of UX problems are not obvious to developers, now ; they will be, at that moment.

1 Like

From my point of view, one can surprisingly quickly get used to workarounds for UI/UX problems. Even more so when there is a production pipeline in place. I believe productions are great to test and expand or modify the functionality of features to a certain extent, but they are not optimal for UX due to the time pressure and a strong focus on one kind of workflow.
In my opinion, the only way to solve those issues is by having people like William Reynish and developers with a strong focus on those parts around. Giving them the opportunity to focus on UX helps a lot more from my point of view than any production. It doesn’t hurt if they tackle UX problems which are relevant for a production, as long as they can focus on UX.


Hi @Metin_Seven,

development is coordinated by very competent people, so I wouldn’t worry.



Blender already has 3-4 filtering levels …
blenderartists, rightclictselect , devtalk … and finally developer.blender_org …
I think the only necessary actions are those that help this filtering flow better.