A couple of weeks ago, Canonical – the people behind the Ubuntu Linux distribution – hosted a talk by the great Scott Jenson, whose work in the user experience of desktop computers goes all the way back to Apple’s System 7. Much of what Jenson said will make sense to anyone who regularly uses a laptop or workstation, which probably includes most of our readership – because in 2025, media production is a rare line of work that remains resistant to being done on a cellphone or tablet.
That creates an interesting conundrum, because until recently, choosing a media workstation was straightforward: something that ran OSX, or, if not, Windows. Until recently, Linux was an option which was only really practical for creative people at companies large enough to have an IT specialist permanently chained, by the ankle, to each creative’s desk. It’s been a while, though, since certain commercial organisations seem to have gone out of their way to make the commercial options less attractive. One company that was very popular in media simply stopped making big workstations, and an operating system vendor decided to occasionally interrupt the process of editing a tragic and visceral war scene with a headline popup about a TV talent show winner.
At the same time, more than one source has suggested that 2025 is The Year of Linux on the Desktop, a slogan that’s been doing the rounds for so long that it demands initial caps. It would be easy to point out that someone’s probably made that claim more or less every year since the late 90s, although recent statistics suggest that the penguin has recently enjoyed a relatively significant – if comparatively tiny – spike in popularity.
Fireproof underwear
In 2010, Linux had perhaps one per cent of the global share of desktop computers (according to the UN’s Information Economy Report in 2012). It then took ten years to get to something like 1.5%. Different sources vary, but since 2020, that number has probably more than doubled. A lot of that is down to Valve’s vast work on Proton making it sort of mostly possible to play a subset of Windows games on Linux. Still, that’s a large inflection on a graph. That’s 100% more in five years, as opposed to 50% in ten.
It’s tempting to ask, in that context, whether 2025 might finally, really be the (caps at the ready) Year of Linux on the Desktop. The thing is, almost everyone reading this probably knows the answer: either you are already a Linux user, in which case the answer is clearly yes, or the answer is still no, just like it has been for the last twenty-plus years, and for much the same reasons as ever.
Addressing this issue requires that we break one of the fundamental rules of online discourse, which is to never, ever discuss politics, religion, or comparative operating systems. So, don your fireproof underwear, because we’re not going to shy from saying that Linux (by which term we’ll henceforth refer to desktop distributions) is both a tinkerer’s paradise and a user’s nightmare.
The fundamental problems are hardly a secret. There are dozens of different distributions, all just compatible enough to be confusing. Hardware and software compatibility varies across distributions and over time as volunteer software engineers chase each others’ changes. Tedious but necessary grunt work around bug fixing and user experience is often overlooked in favour of more creative endeavours. Because most Linux users are technical people, attention goes preferentially to technical people, which is why there are lots of text editors and few video editors. Documentation is often thin or absent because writing it isn’t fun, and the much-vaunted community support can be outspokenly caustic.
This has been the status quo for at least two decades, during which vast amounts of labour has been dedicated to Linux. Certainly, many new things are available. Hardware support, while still substantially incomplete, can only ever increase over time. Most Linux distributions will now happily accommodate a user who needs the proverbial web, word and email. Games are increasingly possible, though mostly through Linux pretending to be Windows, and, yes, there is software like Resolve which offers native support. Still, there is a huge usability problem that is only resolving very slowly.
The solution is often said to be more funding for more software engineers. After decades of exactly that sort of work, though, it might be worth exploring the idea that the problem is not entirely technical.
Ten Linux enthusiasts
As with any complex issue, there are many complex problems, but it’s no great stretch to suggest that the big one with Linux is fragmentation. Dozens of slightly different versions leads to huge duplication of effort, confusion when moving between versions, and friction when deploying code between distributions. Ask ten Linux enthusiasts which is best and discover that they agree on nothing except that all the others are wrong. Google for information on one version, and look forward to a dozen responses about another.
It can be difficult even to figure out what’s possible. Claims that something can be done on Linux might mean that it worked on some version, at some time, on some collection of hardware. Making it happen on one’s own desktop may involve a lot of time and expertise. There are other issues to confront – the difficulty of persuading open-source software engineers to do tedious, essential things, the notorious prickliness of the userbase, outdated ideals of open-source development which struggle to manage large projects, and much more.
If this sounds like advice to be cautious, it is, although depending on interest level it might be worth following up with some distribution-specific information for people minded to try it.
Regardless, the solution to Linux’s longstanding inability on the desktop is in management, not code. While the userbase prioritises the freedom and creativity of a free-for-all development process it is difficult to see how it’s a software problem. The tragedy is that the underlying engineering is generally solid and there has probably never been a better time than now for free software to steal a march on the commercial world.
A choice is involved. To achieve wide acceptance, the Linux world would have to make a decision, to choose a distro, to pick a team, and stick with it until it succeeds – and on that day, the devil will be driving to work on a Zamboni.