Developers, take note: Why people keep flocking to Linux in 2025 (and it's not just to escape Windows)

[FigBug]

I've released plugins for Linux, and honestly, it's just not worth it. Looking at my download stats for one of my plugins: Windows: 41672 downloads, macOS 13806 downloads and Linux 2311 downloads. The effort just isn't worth it for an extra 4% downloads.

I use JUCE, so the porting effort is minimal, but it's not as simple as just firing up another build machine on GitHub actions and sticking the resulting binary on my website.

Linux is primarily designed as an OS for servers and open source software. Therefore, having a stable binary interface just isn't a priority for the maintainers. It's expected that software will be compiled for your particular version of the distro you use.

If you want to compile once and have your binary run on as many machines as possible, it means picking the oldest version of the distro you want to support and linking as much as possible statically (which is a pain since most Linux libraries aren't designed to be statically linked). Example, your plugin needs to do a web call to unlock, so you link curl, but since your dev machine has to run an old OS, it links libcurl3, now your plugin won't run on new OSes because they have libcurl4. So every time a major system library changes major version numbers, your plugin is broken.

The other issue with building on an old os is you are stuck with an old compiler, which means holding your entire project back on an old c++ version.

Now arm is becoming popular for Linux, it means doing another set of builds. But you can't do fat binaries like macOS, so it means another build machine and another set of downloads. Users will download the wrong one, which means more support issues.

None of these issues are impossible to solve, they just take time. And for me the time isn't worth it for a 4% bump in downloads. That time would better be spent on new features or a new plugin.

Asking for help on Stackoverflow is a waste of time, you get unhelpful answers like "X11 doesn't have a concept of application. Nothing prevents your plugin from running its own event loop on its own X11 connection." and "BTW, many Linux computers (e.g. most web servers, or a RasperryPi) don't have any keyboard or mouse." How is that a helpful answer for somebody writing a plugin, I'm obviously not expecting it to work on a webserver. Maybe these days ChatGPT is a better source of info.

A common solution to these issues is to "just make sure it runs under wine", but wine only supports Direct2D 1.2, and now juce uses Direct2D 1.3 so that's no longer an option.

The only way I would recommend a small company do a Linux version, is if they already have a Linux guy already on the team who wants it to be their pet project. They already know the solutions to the above issues, so they aren't wasting days figuring stuff out. If you're a 1 or 2 man shop, you've already got enough on your plate staying on top of windows, macOS, signing, notarization, au, vst, aax, etc.

[audiojunkie]

I've released plugins for Linux, and honestly, it's just not worth it. Looking at my download stats for one of my plugins: Windows: 41672 downloads, macOS 13806 downloads and Linux 2311 downloads. The effort just isn't worth it for an extra 4% downloads.

You are giving only your personal experiences. I wonder how many downloads the Surge Software Team get for Linux for Surge-XT? And the number of users is growing yearly. It is on track to pass MacOS within the next 5 years, if things continue as they are. As it stands, here at KVRAudio, my latest yearly OS survey seems to indicate that 10% of KVRAudio users are Linux users for audio. I'm sorry that you feel this way about Linux. I've always appreciated your work.

Now arm is becoming popular for Linux, it means doing another set of builds. But you can't do fat binaries like macOS, so it means another build machine and another set of downloads. Users will download the wrong one, which means more support issues.

The same can be said about Windows and its new ARM version as well. With MacOS, there's still Intel that developers to a lesser extent are supporting to go along with the ARM based M-series chips. I'd say that this is the way of the world these days.

None of these issues are impossible to solve, they just take time. And for me the time isn't worth it for a 4% bump in downloads. That time would better be spent on new features or a new plugin.

I understand that this is your personal experience, but it may not reflect all developers' opinions. The numbers of Linux users have been growing at a rapid pace yearly, and the OS surveys and growth reports all indicate a high optimism for the coming years. I'm certain things will get better, year by year. Remember, Linux has only recently (in the last 5 years) had so many advances that the average non-tech savvy musicians can even use it. That's progress.

Asking for help on Stackoverflow is a waste of time,...

\

I think you are looking in the wrong places for help. Search actual places where plugins are developed. Try the JUCE forums. Try GitHub discussions section in each software repository where other software plugins, frameworks, and libraries are provided. Try contacting developers that write the programs directly. Try the various discord developers sites, like the Surge Team's discord. @Baconpaul and the gang is always willing to help. Those are the places I've learned what I know about these things--not Stackoverflow--and I'm not even a developer. The best thing for learning I ever did was to join Github and get involved in QA testing for programs I was interested in.

...and "BTW, many Linux computers (e.g. most web servers, or a RasperryPi) don't have any keyboard or mouse." How is that a helpful answer for somebody writing a plugin, I'm obviously not expecting it to work on a webserver.

If I'm not mistaken, those devices are not counted in the "Desktop Linux" numbers. And no one is going to expect you to write a plugin that doesn't support a keyboard or mouse.....unless you are referring to touch screen. But even touch screen is a different case altogether, since you are less likely to see them in a server environment. A lot of what you are referring to are IoT devices, and I don't think anyone is expecting to run a DAW on a refrigerator or a home control center. Just focus on the desktop linux devices with keyboard and mouse.

In fact, this one almost feels like you are "looking" for additional reasons to hate on Linux. I don't see this as a legitimate concern.

A common solution to these issues is to "just make sure it runs under wine", but wine only supports Direct2D 1.2, and now juce uses Direct2D 1.3 so that's no longer an option.

A common and easy solution is to simply fall back to the software renderer. This shouldn't be a problem if you aren't using JUCE 8+. But if you are, the workaround is fairly easy.

The only way I would recommend a small company do a Linux version, is if they already have a Linux guy already on the team who wants it to be their pet project. They already know the solutions to the above issues, so they aren't wasting days figuring stuff out. If you're a 1 or 2 man shop, you've already got enough on your plate staying on top of windows, macOS, signing, notarization, au, vst, aax, etc.

This is a fair assessment. My question for you, is would your stance change as Linux adoption and usage for pro audio changed? What would you say if Linux overtook MacOS in pro audio popularity as it has in the Video Game realm?

Over all, it all comes down to your opinion. You've been an awesome developer for the Linux realm, and I can't take that away from you. It sounds like you are really bitter about Linux right now. My hope for you is that you see the success you deserve from the Linux side of things, so that you'll someday change your stance.

I mean no ill will with my response. I hope you'll recognize this when you read my response.

Best wishes!

[FigBug]

I've released plugins for Linux, and honestly, it's just not worth it. Looking at my download stats for one of my plugins: Windows: 41672 downloads, macOS 13806 downloads and Linux 2311 downloads. The effort just isn't worth it for an extra 4% downloads.

You are giving only your personal experiences. I wonder how many downloads the Surge Software Team get for Linux for Surge-XT? And the number of users is growing yearly. It is on track to pass MacOS within the next 5 years, if things continue as they are. As it stands, here at KVRAudio, my latest yearly OS survey seems to indicate that 10% of KVRAudio users are Linux users for audio. I'm sorry that you feel this way about Linux. I've always appreciated your work.

Last time stats were posted for Surge it was under 5%. This obviously doesn't count anybody that built it from source or installed it via their package manager. But the numbers aren't huge.

Now arm is becoming popular for Linux, it means doing another set of builds. But you can't do fat binaries like macOS, so it means another build machine and another set of downloads. Users will download the wrong one, which means more support issues.

The same can be said about Windows and its new ARM version as well. With MacOS, there's still Intel that developers to a lesser extent are supporting to go along with the ARM based M-series chips. I'd say that this is the way of the world these days.

True, and support for VST plugins on Windows ARM is terrible. There are almost no native arm plugins (Surge is one of the few). I think I've had one customer request. macOS will be dropping Intel support with the next OS release, and I assume we will be dropping intel support a year or two after that. My stats say only 16% of mac users are still on Intel.

None of these issues are impossible to solve, they just take time. And for me the time isn't worth it for a 4% bump in downloads. That time would better be spent on new features or a new plugin.

I understand that this is your personal experience, but it may not reflect all developers' opinions. The numbers of Linux users have been growing at a rapid pace yearly, and the OS surveys and growth reports all indicate a high optimism for the coming years. I'm certain things will get better, year by year. Remember, Linux has only recently (in the last 5 years) had so many advances that the average non-tech savvy musicians can even use it. That's progress.

Hopefully some other developers with Linux native plugins will chime in on this thread.

Asking for help on Stackoverflow is a waste of time,...

\

I think you are looking in the wrong places for help. Search actual places where plugins are developed. Try the JUCE forums. Try GitHub discussions section in each software repository where other software plugins, frameworks, and libraries are provided. Try contacting developers that write the programs directly. Try the various discord developers sites, like the Surge Team's discord. @Baconpaul and the gang is always willing to help. Those are the places I've learned what I know about these things--not Stackoverflow--and I'm not even a developer. The best thing for learning I ever did was to join Github and get involved in QA testing for programs I was interested in.

I'm very active on the juce forum. As well as https://linuxmusicians.com. There just aren't that many people doing Linux VST plugins, information is hard to come by.

...and "BTW, many Linux computers (e.g. most web servers, or a RasperryPi) don't have any keyboard or mouse." How is that a helpful answer for somebody writing a plugin, I'm obviously not expecting it to work on a webserver.

If I'm not mistaken, those devices are not counted in the "Desktop Linux" numbers. And no one is going to expect you to write a plugin that doesn't support a keyboard or mouse.....unless you are referring to touch screen. But even touch screen is a different case altogether, since you are less likely to see them in a server environment. A lot of what you are referring to are IoT devices, and I don't think anyone is expecting to run a DAW on a refrigerator or a home control center. Just focus on the desktop linux devices with keyboard and mouse.

In fact, this one almost feels like you are "looking" for additional reasons to hate on Linux. I don't see this as a legitimate concern.

The quotes I posted above come from an actual stackoverflow question I posted. I was porting a feature that we have on Windows and macOS to Linux, and I couldn't find a to do it. I still haven't figured it out, so the Linux version has less features. This causes support issues when users see something in a video and then they try it and it doesn't work.

A common solution to these issues is to "just make sure it runs under wine", but wine only supports Direct2D 1.2, and now juce uses Direct2D 1.3 so that's no longer an option.

A common and easy solution is to simply fall back to the software renderer. This shouldn't be a problem if you aren't using JUCE 8+. But if you are, the workaround is fairly easy.

Unfortunately falling back to the software render no longer works. As Tom from JUCE says:

That’s where we got to. Even with the fonts installed via winetricks there’s a lot of other things going wrong and we’ve tried a few different ways of falling back to the software renderer without any success.

Ultimately, making this work looks like a big time investment and there are many other different areas of JUCE where that investment would be much more valuable, so we’re not going to investigate any further

You can read the full thread here. The options are either stay on juce7 or maintain a juce8 fork.
https://forum.juce.com/t/juce8-direct2d ... dge/64298/

The only way I would recommend a small company do a Linux version, is if they already have a Linux guy already on the team who wants it to be their pet project. They already know the solutions to the above issues, so they aren't wasting days figuring stuff out. If you're a 1 or 2 man shop, you've already got enough on your plate staying on top of windows, macOS, signing, notarization, au, vst, aax, etc.

This is a fair assessment. My question for you, is would your stance change as Linux adoption and usage for pro audio changed? What would you say if Linux overtook MacOS in pro audio popularity as it has in the Video Game realm?

Over all, it all comes down to your opinion. You've been an awesome developer for the Linux realm, and I can't take that away from you. It sounds like you are really bitter about Linux right now. My hope for you is that you see the success you deserve from the Linux side of things, so that you'll someday change your stance.

I mean no ill will with my response. I hope you'll recognize this when you read my response.

Best wishes!

I follow the users. I didn't enjoy porting macOS plugins from PowerPC -> x86 -> x86-64 -> Apple Silicon. I didn't enjoy jumping through the notarization or signing hoops. I did it because there are a lot of macOS users who buy plugins. If Linux users start buying plugins in similar numbers, obviously I'll have to revisit my stance.

But currently I'm on a tough place. I've released Linux products that aren't up to the standard of their Windows / macOS versions. Users are annoyed because of missing features, but the features might not be possible because the OSes are just different. And if they are possible, it's not worth the effort for me to invest the time into research to implement. Dropping Linux support will also piss people off.

As Linux grows, the average Linux is changing. They used to be very technical, and will to spend the time to figure our how to get things to work. New Linux users just expect stuff to work.

This is just a warning to other devs thinking about jumping into Linux who think it'll be a gold mine because there is less competition from other devs, it isn't. The amount of good press you'll generate by jumping into Linux is a lot less than the bad press if you decided to drop the plugins in the future. Just be aware, supporting Linux is like getting a puppy, it's a life time commitment.

[uOpt]

I've released plugins for Linux, and honestly, it's just not worth it. Looking at my download stats for one of my plugins: Windows: 41672 downloads, macOS 13806 downloads and Linux 2311 downloads. The effort just isn't worth it for an extra 4% downloads.

That's more Linux than I would have estimated for a commercial plugin.

A common solution to these issues is to "just make sure it runs under wine", but wine only supports Direct2D 1.2, and now juce uses Direct2D 1.3 so that's no longer an option.

This sounds like a problem that could evaporate at any point in time.

I wonder whether Valve is interested in messing with that. Direct2D is probably used by a good number of games on Steam.

[audiojunkie]

@Figbug:

That’s unfortunate. I hadn’t heard that the plan to fall back to the software renderer had failed. Initial attempts had looked promising.

I understand where you are coming from. New Linux users coming to Linux and not being as technically inclined is something that was bound to happen. I guess we’ll have to take the bad with the good—there is an expectation of massive growth over the next 5 years.

At any rate, I appreciate what you do!

[camsr]

I think the base distro matters a lot to any software developed. You could move your favorite linux only software from one distro to another if they stopped supporting something.

Agreed. Pick a popular one, and you'll probably be fine.

GTK2 it is then?

Hehehe!!! I don't think you'll find any "popular" distros still using a depreciated framework.

Just the "legacy" programs that require it then?

[audiojunkie]

I think the base distro matters a lot to any software developed. You could move your favorite linux only software from one distro to another if they stopped supporting something.

Agreed. Pick a popular one, and you'll probably be fine.

GTK2 it is then?

Hehehe!!! I don't think you'll find any "popular" distros still using a depreciated framework.

Just the "legacy" programs that require it then?

If there is a legacy program requiring it, it is most likely abandoned. If you do happen to find a program that uses it, I'd recommend running a distrobox container or something to keep your main OS safe. It is well known that GTK2 has been depreciated, and is no longer maintained--it has reached its End Of Life (EOL). GTK2 applications don't work well with Wayland and are really slow. Any up-to-date system using GTK, will likely have GTK4.

[uOpt]

gtk2 is still needed for the good version of Jamin. The gtk3 version is seriously screwed up.

[Kott]

I think the base distro matters a lot to any software developed. You could move your favorite linux only software from one distro to another if they stopped supporting something.

Agreed. Pick a popular one, and you'll probably be fine.

GTK2 it is then?

Hehehe!!! I don't think you'll find any "popular" distros still using a depreciated framework.

wut? gtk2 is still everywhere afaik, just remember that GIMP v3 was released this year

[Tone2 Synthesizers]

Installed AnduinOS 1.4.1 today. It looked great. Installation went smoothly apart from minor issues.
Installed WINE. It does not run with my old Intel Chipset because of "Haswell Vulcan support is incomplete". Bummer.
Hours wasted: 3

Will try ZorinOS 18 now.

[mystran]

Installed WINE. It does not run with my old Intel Chipset because of "Haswell Vulcan support is incomplete". Bummer.

It's been a long time since I used WINE... but if it has a Vulkan requirement these days and you're running Haswell integrated GPU, then switching distributions likely won't help you 'cos there's a chance that the hardware is simply too old for Vulkan.

Something I also want to point out about smaller distributions is that they tend to have somewhat more narrow focus, where they take some base distribution and then they tweak their pet project on top of that ... but because the teams are often quite small (even if the web-page looks very professional), you might run into issues that are much less common with the "big and boring" distributions.

It's also quite common for smaller distributions to tweak the default desktop and what not, but you can generally put any desktop environment in any distribution with not much effort (eg. when installing Debian, it gives you a choice of about ten or so, but even if other distributions have a "default" it's usually more or less "apt install <xyz>" to pull something else in; I run Cinnamon on my Debian box currently, for no particular reason other than it's "not GNOME Shell" and "not as ugly as XFCE").

[audiojunkie]

I think the base distro matters a lot to any software developed. You could move your favorite linux only software from one distro to another if they stopped supporting something.

Agreed. Pick a popular one, and you'll probably be fine.

GTK2 it is then?

Hehehe!!! I don't think you'll find any "popular" distros still using a depreciated framework.

wut? gtk2 is still everywhere afaik, just remember that GIMP v3 was released this year

Yeah, probably an over exaggeration on my part. But it's still a good idea to avoid it if possible.

[Tone2 Synthesizers]

Installed WINE. It does not run with my old Intel Chipset because of "Haswell Vulcan support is incomplete". Bummer.

It's been a long time since I used WINE... but if it has a Vulkan requirement these days and you're running Haswell integrated GPU, then switching distributions likely won't help you 'cos there's a chance that the hardware is simply too old for Vulkan.

Something I also want to point out about smaller distributions is that they tend to have somewhat more narrow focus, where they take some base distribution and then they tweak their pet project on top of that ... but because the teams are often quite small (even if the web-page looks very professional), you might run into issues that are much less common with the "big and boring" distributions.

It's also quite common for smaller distributions to tweak the default desktop and what not, but you can generally put any desktop environment in any distribution with not much effort (eg. when installing Debian, it gives you a choice of about ten or so, but even if other distributions have a "default" it's usually more or less "apt install <xyz>" to pull something else in; I run Cinnamon on my Debian box currently, for no particular reason other than it's "not GNOME Shell" and "not as ugly as XFCE").

Got the same error message in Zorin 18. Here at least WINE does not crash. But it has bad performace and can't handle alpha blending due to the missing Vulkan drivers for the old Intel chipset.
Did not experience such issues with the old Radeon graphic chipset on my other Zorin laptop.

My conclusion: You can intall Linux to reanimate old computers. But these should not be very old and not include exotic hardware. The lack of driver support can suck.
If you are in a professional environment it is usually more cheap to buy a new computer than to waste hours to make things work in Linux.

[uOpt]

My conclusion: You can intall Linux to reanimate old computers. But these should not be very old and not include exotic hardware. The lack of driver support can suck.
If you are in a professional environment it is usually more cheap to buy a new computer than to waste hours to make things work in Linux.

Haswell is hardly "exotic hardware".

It is very old, though. I doubt any professional environment use it for 3D.

[ghettosynth]

My conclusion: You can intall Linux to reanimate old computers. But these should not be very old and not include exotic hardware. The lack of driver support can suck.
If you are in a professional environment it is usually more cheap to buy a new computer than to waste hours to make things work in Linux.

Haswell is ancient in computer terms. I still have a couple of gen 3/4 machines hanging around, but I'm not trying to push them into places where they're not suited. One is being used as a virtual host for four or five linux servers, it works great for that. If I had to replace it today, I wouldn't buy it, one of the newer N series chips is fine for that purpose and will sip power.

I think the idea that you can turn a decade+ old machine into something similar to a modern machine just by using Linux is something of a myth itself. It's not if you are using it as a server, or as a desktop with a really light window manager.