At the Linux Application Summit (LAS) in Barcelona, GNOME, KDE, and other Linux developers came together to work on making incremental process on the Linux desktop. NextCloud founder and former KDE board member Frank Karlitschek had another, bigger idea: Reclaim the "Year of the Linux desktop" as a real plan rather than a joke.
After all, Karlitschek reminded his audience, in the 90s, when the free software/open-source movement really got traction, no one thought Linux could take over the server. But it did. So, why didn't the Linux desktop, as well? Sure, there were all the business reasons Microsoft dominated the market, but the Linux desktop also suffered from many self-inflicted wounds. And, unfortunately, Karlitschek thinks Linux desktop developers are still indulging in the insanity of doing the same thing over and over again but expecting different results.
Karlitschek, who noted he was a dinosaur and no longer in the Linux desktop per se, said he still wants it to succeed. He thinks there is still a need for a free desktop, but that desktop must be more welcoming to third-party independent software vendors (ISVs). With that, he said, "We can give normal users stuff like Microsoft Word, Photoshop, games, and enterprise applications like Oracle and SAP." Compared to any other end-user operating system, Linux has only a handful of applications.
So, how can Linux do this? Karlitschek had several practical suggestions.
First, there needs to be a central developer portal. This will be where ISVs can find the documentation, tutorials, software development kits (SDKs) and examples they need to get up to speed and build applications. They don't have the time, Karlitschek said, to search for the answers. Adobe, Apple, and Microsoft all do this well, but Linux really doesn't have anything like this.
Next, ISVs need stable application programming interfaces (API)s. "You can actually run 20-year-old Windows applications still on the modern windows. That's crazy, but you can," said Karlitschek. With Linux, you're always needing to recompile links to the newest libraries.
Continuing with this theme, Linux also needs consistent APIs. For example, from his own perspective, the NextCloud has to work with both GNOME's Nautilus and KDE's Dolphin file manager. "Why are the two APIs for basic file management? Why are you asking ISVs to do double the work?" And, of course, those are only two major file managers. There are many more.
One old pain point ISVs face is which toolkit should they use. While that still causes a lot of heartburn for desktop developers, it's no longer a real issue for ISVs. They pick the software toolkit they want, and they make it work. No problem.
Packaging, however, that's still a pain in the rump. True, the rise of Snap and Flatpack, which provide containerized applications that can run on multiple Linux distros, has made it easier than ever for ISVs to deliver their programs on Linux. But why do we have two of these? Why can't we have one?
And, why, for that matter, with the Linux desktop having at most 3% of the market, do we have two major Linux desktop foundations? Karlitschek was very happy to see GNOME and KDE coming together at LAS. But why not join together, or at least work together, under a joint umbrella foundation?
For that matter, isn't it time, since Microsoft is now bringing end-user applications to Linux using Electron, a high-level cross-platform development platform, such as Teams and Visual Studio Code, to encourage Microsoft to bring all of Microsoft Office to Linux? Or help Adobe to bring Photoshop to Linux?
Free software purists may not want it, but ordinary users do. After all, as others at the conference observed, you can now run Office 365 on Linux.
Not everyone at the conference was happy about his end comments. LibreOffice and GIMP supporters weren't pleased with the idea of their chief rivals coming to Linux.
But his main points were well received. The majority of the Linux desktop developers at the conference agreed that they need to unite their fragments into a whole and that every effort should be made to make life easier for ISVs to port their applications to Linux.
Who knows. If Karlitschek's ideas are put into practice, maybe there will be a true "Year of the Linux desktop." After all, it's clear that not everyone is happy about having to move from Windows 7 to Windows 10 and Microsoft itself is intending on replacing its traditional Windows desktop with Windows as a Service. The time for a mass-market Linux desktop may yet come.