This should make it possible for me to write some seriously flexible GPU-accelerated code now. But the issue is that if I release an addon that relies on this, only 5.3+ will be supported, and this excludes the current LTS version.
I dont like the idea of not supporting a currently active LTS, so I’m wondering what you guys think?
Should I wait a year for the next LTS before releasing a tool that uses it?
Release it for blender 5.3 and just not support 5.2-LTS?
Jurry rig some spaghetti so the same addon works on 5.2, even if poorly?
I personally design all my add-ons to support Blender versions as far back as reasonable, usually to 3.0 if I can manage it. This usually just requires a quick check here and there.
That said, if your add-on relies on something that is only available or reasonably feasible in 5.3+, then just marketadvertise it as such.
Remember, there are still people out there using the old 2.x series. No point in targeting people who can’t use it. There are already plenty of people downloading and using the 5.3 pre-releases.
At the risk of being a wet blanket, the current reality of Blender add-ons is that people aren’t buying them. We live in a world where any given free chatbot can make or duplicate any given addon. The amount of new paid addons being sold on BlenderMarket is a flood. No one is buying them, everyone is making them.
Is your motivation to make a tool that benefits your workflow? Make it for whatever version you use. Is your goal to make a killer profit? Find a different way
Well, we don’t have the results to the 2026 Blender Survey yet, but the 2025 one showed that most people don’t use the latest version of Blender, however, a fair percentage do. So it just depends on what you want to do really.
Me personally, I always use the latest stable version of Blender because, unlike other software, Blender releases actually give important new features. But I also don’t generally use addons (too likely to break).
well, I was just about to start making a GLSL parser/transpiler to convert struct defintions into imageLoad/Store with texel offsets so that I can make a wrapped texture pretend to be a SSBO, then it would work on 5.2 too, and that was my plan from the start until the pull request dropped.
Now I don’t wanna do any of that because its a ugly solution compared to just straight-up using the new SSBOs.
Yeah it seems to me as well it’s not much of a question. If it’s only possible in version 5.3, then the question is only if you wait a year. Nothing else changes. So you either loose a year of feedback and potential sales or you don’t the way I see it.
Yea, then that settles it.
This will be the first time I target a feature that isn’t even on a stable release at the time I’m working on it. Its a weird feeling, since I’m used to test and make a tool works on at least a couple of older blender versions.
Apart from what joseph said about Blender add-ons and selling them, which makes sense to me, I don’t really see a problem with this. New features are new features, I think it’s reasonable that if there is a technical reason then that’s it.
I think if I were you, I would wait before doing all this
Sounds like a lot of work, maybe it makes sense to see if people need this. I think they ask for stuff if they need it. Might make sense to wait for at least a few user requests before investing your time into that. At least that’s what I think. Not that I know anything about selling add-ons. I just share them… Just thinking out loud.
Agreed. If something needs to be rigged to sort of approximate what the new feature does nicely, and especially if it requires a whole bunch of work to do it, then wait for people to start asking for older-version versions en masse.
The whole point of new versions is new capabilities — one of which your plug-in provides, right?
It is easy to overthink it. We all fall into that trap sometimes. Glad you’re satisfied.
I had go with a hybrid approach: release for Blender 5.3 with full support, while keeping 5.2 LTS compatibility only if it can be handled cleanly through a separate fallback. I wouldn’t delay a useful tool for a whole year or introduce fragile workarounds just to maintain compatibility. Clear version requirements and a reliable experience are better than supporting every version poorly.
..but then there are two versions to be developed and maintained. Also when the 5.2-verson is inferior to the 5.3+ version then users who tries this will say: “nay it isn’t worth” and so will be no future customers..
..so your last sentences seems to contracdict you first two.
Well, I suspect it’s the case for simple stuff but there is still super advanced things that can hardly be recreated.
In the meantime, there is so much addons and assets nowadays I’m wondering how long this could stay sustainable… In the meantime recently I’ve seen a bunch of commercial projects that relies heavily on addons or assets … It’s a double edged sword because it adds a lot of overheat to projects. One I have in mind was basically having super heavy scenes because it uses a mix of ivy generators, rain generators, VDB for smoke effects, probably stuff like geo-scatter for vegetation and I suspect many more.
This worked fine at some point but this also become a nightmare to render and control all those side elements. While it’s possible to ask a TA to do all those simplified setups in a unified way and avoid unnecessary complexity.
Initially assets and plugin were a side thing as basically artists where selling their tools or their work . After a while this became a job on it’s own but I’m really wondering how will that evolve in a few years from now … Maybe the trend will wear out and eventually that will leave place for this buisiness to rise from ashes again…
As for the OP’s question, I would stick to 5.3, but that depends on what you’d like to do…
People using LTS are generally under long productions and they are less likely to plug a new tool in the middle of that… Most users should use the latest version unless they are forced to stick to one. And probably in a few months or a year the problem will solve by itself and you won’t have to manage complex code to support every possible blender version. Anyway, maybe it’s better to ask your user base and seek input from potential clients…
So you either loose a year of feedback and potential sales or you don’t the way I see it.
..