Let”s look at the most common arguments and mistakes on their part. But only about Linux, and a full-fledged desktop one! And a little about Open Source in general. You can talk for a very long time how bad everything is with this or that non-Linux-based OS or program, but Linux will not get better (or worse) from this. You can convince for just as long that “everything works for me”, or blame the opponent for incompetence, but this will not change anything either. And you can finally admit that there are problems and that the average user is not able to cope with them.
Argument 1: Linux is everywhere
Or, more simply, Linux has already won, so you can sit still, not worry and enjoy the victory. Here they usually remember embedded solutions or, in general, some ready-made products (more often they talk about network products), as well as Android and supercomputers with servers.
It”s easy to make sure that these claims are true – the 2017 Linux Kernel Development Report cites the following figures: 90% of desktop applications in public clouds run Linux-based operating systems, this type of OS is 62% in the embedded market, and in the supercomputer market – all 99% (now, of course, even more). Finally, 82% of smartphones in the world use the Linux kernel, as do 9 of the top 10 public clouds.
How all this relates to desktop distributions, in the discussion of which such arguments pop up, is not very clear.
Let”s sort it out one by one. For embedded solutions, the Linux kernel is indeed used, but only the kernel. Even the usual minimal environment may not exist; at best, busybox will be attached. And the kernels themselves in this case are also quite deeply customized. But these are all details, the important thing is that the user ultimately doesn”t care what”s under the hood. He needs a certain set of functions and an interface to work with them.
With Android, the story is exactly the same. At the core, of course, is also the Linux kernel, but the vast majority of applications interact primarily with the Android Runtime (ART), and not with the kernel. Actually, this very core also differs in many respects from the main branch, and the same ART has already been ported to Google Fuchsia. So if at some point developers decide to move from the Linux kernel to any other platform, the end user will probably not care.
Everything seems to be clear with supercomputers – the most powerful systems in the world run Linux, and here we are talking about the OS, and not just the kernel. Servers are viewed in the context of the infrastructure that somehow provides our life, directly or indirectly. Everything, of course, is so, but the question is the same: how does this all relate to desktop distributions? Especially if you remember that behind all of the above solutions there is a whole staff of system administrators and developers.
Argument 2: Linux is free
It would be more correct to say that Linux is more profitable or, say, cheaper than other solutions. But there is no sublime altruism and high spirituality here. If you are not paying right now to download an iso image and use some distribution kit, then this does not mean that you are not paying at all. At the level of a user or even a company, the transition to Linux still translates into time spent on mastering and maintaining the system, and time is money, no matter how trite it may sound. And this, in my opinion, is one of the key constraints.
If you look more broadly and philosophically, it is time to admit that Linux in all its forms is quite commercialized. Somewhere at the turn of the century, an important turning point occurred. If earlier the main users of Linux were developers, often from the academic environment, who themselves could add or rewrite something in the system, then later these two groups began to coincide less and less.
Many major and independent, that is, sufficiently different from others, modern distributions either have free and commercial (this also includes support) versions, or are directly or indirectly sponsored by other companies, institutions, the state, and also individuals. For example: in 2018, only within the framework of SPI, Debian had $ 335 thousand, and Arch Linux – $ 294 thousand. And this is the norm for seemingly free projects. In particular, the Document Foundation (LibreOffice) costs € 700-800 thousand a year, OSI spends several hundred thousand dollars, and the Linux Foundation, a key organization in the Linux world, is rapidly increasing its turnover – in 2017 it received more than $ 80 million.
It is, of course, a penny, but there are other ways of support, including the provision of equipment, hosting, traffic, event space, and so on. The norm is the hiring of individual developers to develop and support specific functionality, and business developers, and the presence of developers on the staff who are solely engaged in the development of Linux or other free projects.