Is Linux A Real Option For Music?

[TechHaus]

Apple sucks for normalizing app stores taking so much money from a transaction. Kind of evil all of the money they sucked up.

[audiojunkie]

Microsoft Corp. seems not able of stopping itself from putting all that tons of telemetry-related junk in Windows (and Apple is even worse)

Public opinion does not agree with that sentiment:
Screenshot_20260605_211820_Mastodon.jpg

Wow! And here I thought Lamashtu was a cinch for the win! Looks like a legit poll to me!

[audiojunkie]

Getting a clean and high-end, truly professional Windows OS installation is possible still today, but it's getting really harder and harder and time-consuming (ReviOS).

I have a dream...

Since it's evident we are living in the time of barbarians, culturally speaking (with no offense for the real Barbarians) and Microsoft Corp. seems not able of stopping itself from putting all that tons of telemetry-related junk in Windows (and Apple is even worse), I really hope in an Open Source initiative (a really huge and solid project involving 1000s of devs) putting together Linux + Wine + etc. (ReactOS?) so we can finally have a free, professional (and easy to install and set up) OS based on Linux, but (extremely important) capable of carrying the whole heritage of Windows Softwares/DAWs/Plugins/etc. (from XP to contemporary 10/11 and further).

And say a final goodbye to Microsoft and Apple and their anti-computerscience behaviour.

As someone who has followed ReactOS since its beginnings....... You'll be waiting a long time for this wish. It's really not that bad to port to Linux, according to many of the developers I've spoken with. The key is not to try to support all distros, but to support one old one: Debian. Libraries are backwards compatible. So if you compile your binaries in an old version of Debian, it should be compatible with most all distros. For example, compile to Debian 12 (current version is Debian 13). Create one .Deb and a Zip of the binaries. Debian, Ubuntu and all derivatives can use the .DEB, which covers the most popular distro families. Everyone else can copy the binaries to from the Zip file to the proper folders. It should work for most everyone--quick, clean and mostly problem free.

[xhunaudio]

Yes, it was just to describe the problem in 2 words.

ReactOS is proceeding very (too much?) slowly in order to be sure to use only 100% legal, original, reverse engineered code.

Windows XP/7 were top, workstation-class OSes out-of-the-box. With 10/11 things changed. I can't figure out why I have to spend all that time making 10/11 great again with ReviOS (great initiative). Simplicity should be something out-of-the-box of a product, as it was (once upon a time) with systems like W XP, W 7, IRIX, Solaris, etc.

I would like to see ( and support ! ) an official Microsoft Windows including just the Kernel, a bunch of System Utilities and a few simple common-use Apps (like Paint). Period. An OS core of just a few 100s of megabytes.

Several ex-Microsoft researchers/engineers stated Windows today misses a pro-mode without all that online activator/telemetry/cortana/agents/AI tons of junk.

So my dream : if Microsoft don't want to make Windows great anymore, it would be great to have a real alternative...

In that sense, long live Linux + Wine - or ReactOS, etc. - any kind of effective open source project on that direction !

It would be great to have a free (all meanings of "free"), offline, private OS for workstation-class purposes out-of-the-box, with a deep and complete compatibility and support for all the priceless Windows softwares of the past decades... (32-bit and 64-bit). A real computer-science oriented OS (like with XP/7), outside the current "feed the AI with plebeians" purpose/mechanics.

I don't want to say anything about Apple, I do not consider it a computer company. You can't disable anything from macOS, but I think fanboys are completely fine with that - and with the very colourful Apple Store etc. And same for Google, obviously.

[TechHaus]

Sad moment for me after talking so much junk on here about how solid my Linux system is...had to hard reset after it wouldn't wake from sleep just now!

On KDE's "Power and Battery" screen I am activating "Manually Block Sleep and Screen Locking" for a while.

I don't feel like figuring out what caused this.

[camsr]

Sad moment for me after talking so much junk on here about how solid my Linux system is...had to hard reset after it wouldn't wake from sleep just now!

On KDE's "Power and Battery" screen I am activating "Manually Block Sleep and Screen Locking" for a while.

I don't feel like figuring out what caused this.

Check your BIOS for any relevant power settings.

[Tiles]

Getting a clean and high-end, truly professional Windows OS installation is possible still today, but it's getting really harder and harder and time-consuming (ReviOS).

I have a dream...

Since it's evident we are living in the time of barbarians, culturally speaking (with no offense for the real Barbarians) and Microsoft Corp. seems not able of stopping itself from putting all that tons of telemetry-related junk in Windows (and Apple is even worse), I really hope in an Open Source initiative (a really huge and solid project involving 1000s of devs) putting together Linux + Wine + etc. (ReactOS?) so we can finally have a free, professional (and easy to install and set up) OS based on Linux, but (extremely important) capable of carrying the whole heritage of Windows Softwares/DAWs/Plugins/etc. (from XP to contemporary 10/11 and further).

And say a final goodbye to Microsoft and Apple and their anti-computerscience behaviour.

As someone who has followed ReactOS since its beginnings....... You'll be waiting a long time for this wish. It's really not that bad to port to Linux, according to many of the developers I've spoken with. The key is not to try to support all distros, but to support one old one: Debian. Libraries are backwards compatible. So if you compile your binaries in an old version of Debian, it should be compatible with most all distros. For example, compile to Debian 12 (current version is Debian 13). Create one .Deb and a Zip of the binaries. Debian, Ubuntu and all derivatives can use the .DEB, which covers the most popular distro families. Everyone else can copy the binaries to from the Zip file to the proper folders. It should work for most everyone--quick, clean and mostly problem free.

This paints an inaccurate picture of how Linux works and what it is.

The claim “Porting to Linux is not that hard” is simply not true. I would suggest learning how to develop and working on real projects before making such statements. As soon as you depend on Windows-specific APIs, DirectX, plugin ecosystems (VSTs), or proprietary audio/graphics stacks, you are usually looking at a partial or even complete rewrite, not a port.

The claim that you can just target an older Debian version and get broadly compatible binaries across all Linux distributions is also misleading. Linux does not provide a stable cross-distribution ABI in that sense. Differences in glibc, system libraries, audio stacks (ALSA/JACK/PipeWire), graphics layers (X11/Wayland), and driver environments regularly break that assumption. That is exactly why Flatpak, Snap, and AppImage exist. And even those are no panacea. Canonical, for instance, is openly fighting against Flatpak.

Likewise, “ship a .deb plus a zip and you’re done” is not how it works. That only works on Debian-based systems. And even then, you are not “done” as you are on Windows, where an .exe runs just fine for the next fifteen years. The next point upgrade of your primary target distribution might already require you to repeat the whole procedure. And that is for just one distribution. There are over 600 distributions with conflicting package managers and ecosystems. Even a specific version can break compatibility assumptions, and even a point release can break software compatibility.

This is not a matter of opinion, and these points have already been discussed several times.

[xhunaudio]

It would be a very hard task indeed - and I'm not strictly talking about modifications of Linux itself at any cost.

It's also worth mentioning - things way more complex and difficoult have already been accomplished, also as open-source projects. With a solid team of passionate and talented software architects (also 100s or 1000s), such new OS could be made, definetly.

Today, the reason of a potential migration to Linux (and this is why this same kvraudio topic/discussion/thread exists) is because Linux's appeal (also as a Kernel) resides in its being free and open-source - this is why people put attention to it.

So, generally speaking - dreaming of a whole new open-source OS does not strictly require an existing Linux (or ReactOS, etc.) Kernel, or a modification of a current Linux distro.

This new "Fair OS" (let's call it this way) could be built starting from the ground up, also with - let's say - **intense** reverse-engineering practices. This means an open-source project is way more transparent and "fluid" compared to a private Company project : trying to stop a distributed, open-source project is the same as trying to stop the ocean with hands. If there're solid, ethical reasons and ethical contexts, a decentralized open-source project has the power of an earthquake. I can't be more specific at this level.

I think today there are huge solid premises to consider this "Fair OS" project to soar. Real Computer Scientists ethically *deserve* a solid OS, they *deserve* full Softwares preservation across decades (as a preservation of a form of Art) through Hardware Emulation also at Driver level, they *deserve* to stay away from the jail/cage of the current AI-training dystopia or dysto-upidity mechanics.

Too long to explain it in detail, but I hope my simple words may inspire projects from people of good will, instead of just cause the usual insults from the usual stupids.

[Tiles]

I think the biggest issue with this Fair OS vision is that it underestimates the importance of the ecosystem.

Building a kernel is hard. Building a complete operating system is harder. Building an ecosystem that can compete with decades of accumulated software, drivers, tools, documentation, standards, community and developer support is on an entirely different scale.

Ironically, ReactOS is probably one of the best examples of why this is so difficult. The project started in the 1990s with the very specific goal of recreating Windows compatibility. It did not even try to invent a completely new platform. Instead, it focused on reproducing an existing one with a known API and software ecosystem.

Yet after decades of development by many talented contributors, it is still not a practical replacement for Windows. Not because the developers are incompetent, but because operating systems are enormously complex and compatibility is a moving target.

If reproducing a single existing platform has proven to be a multi-decade effort, then creating an entirely new operating system, a new driver model, a new software ecosystem, and a new hardware compatibility layer from scratch would be an even larger undertaking.

That's why I believe the challenge is not primarily technical brilliance. It is scale. The ecosystem is the hard part, and ReactOS demonstrates exactly how difficult that problem really is.

[xhunaudio]

No doubt it would be a complex and vast challenge indeed. Also just writing down the first specs would be difficoult. It does not have to be ready for tomorrow afternoon, anyway. ReactOS was started decades ago as a private (non-commercial) project, it resided on a specific website, etc. If Microsoft closed that website 20 years ago, ReactOS would not survive that move (this is why ReactOS had very slow progresses : in order to prove each single line of code was "manually" reverse enginnered and not copied in any way from the source of Windows). Today instead, things could be speeded-up a lot and open-source projects diffusion is way more un-controllable for a single company (also if that company owns gith*b). A project can be forked and forked again on other places with other names.

Also, take the example of the recently released Sw*y*d : it is a 100% legal and fair project, it works without including the "original, vintage files". Then, if happened that you personally own the "original, vintage hardware" and you manually dump the "original, vintage files", maybe you can import that files too and have a complete replica of that hardware.(Don't want to be more specific).

Why do not take this approach for OSes too? It would speed up things a lot : if a 3rd party wants to make a "brick" or module (see below what a brick is) for FairOS with the purpose of "vintage system calls" using an emulator and the dumped "original, vintage files", who can stop it? Anyone thinks this would be a non-ethical practice? Really? Instead - 1 example of 1000s - it's deeply unethical to spend minutes of your life fighting vs your UEFI in order to make it permit you to install a Linux distro on your system... but I bet there's tons of idiots out there saying "that's for our security and safeness, thank you".

Anyway, since this is just a dream for now, imagine : a free OS with a single Kernel at the center of a set of modular bricks each targeting a different aspect of the OS. Some 3rd party could make a hardware emulation brick, others may be interested in emulation of vintage system calls of libraries and make a brick on this purpose, etc. Once a brick is completed and working, it will be compatible forever with subsequent iterations of this free FairOS. Modular nature. No obsolescence. Minimal impact on performances could be reached too. Nobody has to throw away the old wheel, then re-invent the same wheel, just to have the possibility to re-sell the same working wheel twice every 10 years.

For now, it's just a dream. But maybe we already had this OS, if in the passed decades the decisional power was *entirely* in the hands of Computer-Specialists instead of money-specialists.

This is why open-source is ideal for this kind of stuff. It stays away from money-only (or worse) mechanics.

[Tiles]

I think we’re actually agreeing on the “open source makes things more distributed and harder to centrally shut down” part. That’s true in principle.

Where I think the conclusion goes off track is the assumption that this automatically makes a full new OS ecosystem feasible or significantly easier.

Linux is probably the best counterexample here, not a support case for it.

Linux is completely open, massively distributed, forkable, and backed by decades of global contributions. It already is the “maximal version” of what you are describing in terms of openness and decentralization.

And yet it still hasn’t replaced Windows or macOS on the desktop in any meaningful way. Not because of control or suppression, but because of ecosystem gravity: drivers, proprietary software, vendor incentives, compatibility, user expectations, and commercial tooling all cluster around existing platforms.

Forkability doesn’t really solve that. You can fork code, but you can’t fork hardware vendor support, application ecosystems, or developer attention at scale.

Even ReactOS again is instructive here. Its limitation was never “central control of a website.” It was the sheer complexity of reimplementing a moving, undocumented target plus the surrounding ecosystem dependencies. Distribution doesn’t change that core difficulty.

So I would separate two ideas:

Open source absolutely helps resilience and transparency. But it does not fundamentally reduce the cost of building and sustaining a full operating system ecosystem from scratch. That cost is still dominated by hardware compatibility, application availability, and long-term maintenance across decades.

That’s why even extremely open systems still converge around a few dominant platforms. Not because alternatives are “blocked”, but because ecosystems naturally concentrate.

[xhunaudio]

Well, I hope there will be attempts in that direction, anyway.

Although, a good "pro-mode" Windows 12 release would be more than enough in order to improve things a lot. They can put all the junk in "smartphones OSes" if they can't avoid to steal data in any way...

I also remember Windows 12 was planned to be "modular", without further specifications from Microsoft.

[Tiles]

Yeah. I would already be happy when these folks would stop the AIfication of the OS. I am AI addict, but that's the completely wrong direction.

[knob_monster]

this is why more devs are moving towards web standards for utility software.

building native desktop apps today means dealing with expensive code signing certificates, apple developer taxes, windows smartscreen warnings, and broken linux dependencies.

[Tiles]

I wish it were that easy. The browser is part of the ecosystem, not a replacement for it. Web apps don't replace the platform underneath.

Don't forget that a browser is an additional abstraction layer with its own constraints: sandboxing, memory overhead, API limitations, reduced low-level hardware access, and an extra runtime between application and OS.

So you're not removing complexity. You're adding another layer on top of an existing one and trading one set of constraints for another.