The Dream of The Linux Desktop

In a (large) video meeting at virtual FOSDEM, someone joked that maybe this will finally be the “year of the Linux desktop”. One of the oldest jokes in the community. Everyone sniggered. But the joke’s on them: it already happened, like a thief in the night.

Virtual Brussels, home of Virtual FOSDEM

While we were all waiting for the open source community to topple Microsoft’s desktop monopoly by replacing the locally-installed operating system, we missed the real revolution. Sure, there’s still plenty of money in both operating systems and in desktop apps. Microsoft (and indeed Apple) will be milking that legacy monopoly for a good while even if they have started hollowing out themselves — on Linux. It’s certainly been the target of competitive attention from open source software from the beginning; indeed, the open source productivity suite family now epitomised by LibreOffice has over its long history done an effective job in opening up that part of Microsoft’s monopoly (even if mostly by triggering the creation of ODF).

Desktops are there in your face all day and there’s this one app you love, so it’s natural to treat them as a priority. But most enterprise expenditure doesn’t happen on the desktop. Maybe it’s software Stockholm Syndrome making us all love our captor, but the focus on desktop applications, coupled with the idealistic expectation that Windows will be displaced, has led many to overlook or even dismiss the reality that Linux actually has taken over the desktop already.

Presto, Changeo

That’s in the browser. Think about it: When did a new process or service you wanted to use last come as an application download? When it did, what actually was that application? An increasing number of desktop applications are just containers for Web apps, or shims for web back-ends. The real powerhouse behind those apps is usually Linux, accessed over the Internet, along with other elements of the modern Internet server stack. Moreover, the servers themselves are Linux all the way down, from containers to hypervisors to controllers.

In a very real sense, the applications many use daily for email, documents, presentations, and more are in fact Linux desktop applications. A fanatical obsession with replacing Windows made for interesting discussion, but while that debate was happening, all the work on the desktop moved inside the browser window and the software powering it onto Linux. The Linux desktop is called “the browser”.

In turn, that desktop revolution has fuelled — and been fuelled by — Linux in portable devices. In that space, Linux is definitely winning globally, both by powering multiple device platforms such as Android and Kindle and by powering many of the applications found on those devices, not least using PWAs and web back-ends again. A developer can take the same Linux-powered architecture and use it for both desktop/browser apps and for device-specific mobile apps. The result? Linux is everywhere, even on iOS; it’s just not advertised.

Things are going a step further. Plenty of people (myself included) use a laptop running a browser directly on a Linux kernel, with a container system that lets it host Android, Linux and even Windows apps. It’s centrally maintained and secured and thus the hardware is drop-in replaceable in seconds. In many ways Google’s ChromeOS is the perfect conclusion of the Linux desktop journey. If only it wasn’t controlled by Google surveilling my every click for search supremacy.

It was natural to assume the wedge to displace Microsoft’s desktop monopoly would be something that did the same thing. Futurists have long made their predictions by describing the present wearing gold Spandex. When it comes along, the new reality often looks different from the future we expected. So we still have no flying cars, food still doesn’t come in pill form — and the Linux desktop is actually running in your browser or behind what looks like your app. The year of the Linux desktop came long ago and we missed it.

Yes, we did it folks.

But in another sense it hasn’t happened, and looks less likely than ever. The reason so many people continue to “look to the east at dawn on the fifth day” is the dream of Linux on the desktop was never really about the software itself; it was about the freedom it embodied. Open source software depends on the presence of the freedoms to use, improve and share the software and its source code, and for many of us the fact we controlled the software ourselves was the point, even if we expressed it as a longing for the infinite choice (and complexity) of Debian.

The dream of “the year of the Linux desktop” was actually the dream of software sovereignty. Of an end to the “upgrade treadmill”, of serfdom to corporate lock-in masters, to a system that works just the way we want it. Today that is further away than ever; in fact the Linux desktop as it has been realised brings us less control over our data, over our privacy, over the works we create, than we had even when the computer in front of us was running Windows in 2005.

Adapted with saudade from an InfoWorld article I wrote in 2013