On the useless hope of FLOSS: why the average Joe doesn't care about Linux (and never will)

Aug 19th, 2026 - scudo

In the past 20 years, in the FLOSS world there has been this theoretical magical target of "the year of the Linux desktop", this idealized world where everyone (or at least most people that use computers) is freed from the chains of proprietary operating systems and runs a completely free and open source system.

I have been around the open source world for 5 years at this point (I started using Linux in 2021), and the more time passes, the more improvements the Linux desktop sees, the more I am convinced that "the year of the Linux desktop" is bullshit and most likely a nonsensical and irrealistic target to hit, unless a gigantic and actually real paradigm shift happens in computing. Which I don't see coming, but I'd be glad to be wrong about this whole post, frankly.

The "Bob in accounting" problem

The root issue is, taking inspiration from CRD, that Bob in accounting does not care about Linux, or open source software; the only thing Bob in accounting cares about is whether or not the computer works so he can do his job.

Most people (except tech-savy users) do not care about what platform their applications run on: they do not give a crap whether the programs that they want to use run on Windows or Linux or macOS, they care about doing things with the computer. Most people don't even know (or care about) what an operating system even is, because they are not paid to give a shit about it. People don't know they are learning Windows, they think they are learning "the computer".

And this is not a "Windows has a monopoly" issue; Windows just happens to be the platform that most applications that people want to use are written for. If Linux had that, people would be complaining about "Linux has a monopoly and you should use BSD (or whatever Unix-like OS is popular at that moment)".

The platform does not matter: the Commodore case study

The fact that a platform is not defined by its technical capabilities but by how much software people want to use the platform has is something that should be painfully obvious to anyone that knows something about the now resuscitated Commodore computer brand; they released several computers in the mid 1980s that by all accounts should have been massive successes in the marketplace (the Plus/4, the C128 and the Amiga), but all those different and incompatible computers (besides the Amiga) were market failures instead. Why so? Well, let's have a look:

• The Plus/4 had several machines with different hardware configurations so programmers targeted the lowest common denominator so software was lacking for those that bought more expensive machines with more memory; also they should have been cheaper (not my words, Bil Herd's). So the few pieces of software available weren't even good.
• Nobody used the C128 in... well... C128 mode, because most people used it only in C64 mode, for which there was a vastly bigger software library, so no programmers were interested in targeting the C128 specifically, because everybody used it in C64 mode... do you see the infinite loop? (This is the main reason why nobody used the C128's 80 column mode... which was the main reason why the computer was made in the first place)
• The Amiga saved its ass because of Deluxe Paint and video games, but that stuff came late, and the platform had a slow start because of it (see this documentary on the matter)

Ergo, the problem is that there wasn't really anything to do with those machines. So people didn't buy them. The same thing applies to operating systems: why switch to Linux when I already (or only!) know how to use Windows and that is the magic thing that makes my programs go? Why should I bother with a completely new thing when the old thing does what I need to do and the new thing may not even work properly with my programs?

The feature parity problem

No amount of features in the Linux kernel, in the various desktop environments, in all the software packages, is going to make Bob from accounting want to install Linux on his PC, because he does not care and just wants Microsoft Excel to run because that's all he knows about the computer.

There is this mentality that has been around for decades in the Linux world that when we have all the features that Windows has, Linux is going to gain significant market share. This mentality makes no sense to me for several reasons:

1. Windows and Linux operate under completely different paradigms and are developed with different OS design philosophies in mind; they are fundamentally different in basically everything, so it is probably impossible to copy everything over;
2. As I stated before, Bob in accounting does not care about feature parity, and no amount of Wineing is going to make him care;
3. Some of the stuff Windows does is probably still under patents.

We are so intent on having feature parity with Windows that we don't even match some of its objectively better design patterns (see WinSxS) to fix issues the Linux desktop has had for literal decades (especially when talking about side-by-side assemblies to remedy DLL hell as I just talked about).

Conclusion

There isn't really a conclusion per se. The fact is, you can't apply a technical fix to a social problem, and it really seems like we are doing just that.

I really, really hope this post ages like milk, because I would love to see the year of the Linux desktop, eventually.