For those who try to keep their finger on the Linux community’s pulse, 2018 was a surprisingly eventful year. Spread over the last 12 months, we’ve seen various projects in the Linux ecosystem make great strides, as well as suffer their share of stumbles.
All told, the year wrapped up leaving plenty to be optimistic about in the year to come, but there is much more on which we can only speculate. In the interest of offering the clearest lens for a peek into Linux in 2019, here’s a look back at the year gone by for all things Linux.
Ubuntu Sheds Unity but Sees Silver Lining in Cloud
The last ripples from 2017 into 2018 came from Ubuntu’s decision to phase out the Unity desktop and switch its flagship desktop environment to Gnome. Ubuntu’s first image to ship with Gnome was with its October 2017 release of 17.10, but it was something of a trial run. With April’s 18.04, Ubuntu officially unveiled its first Long Term Support (LTS) track to feature Gnome 3.
With an LTS sporting Gnome and holding up to user testing, the countdown clock began on the eventual switch to the Wayland display server, intended to take over for the aging Xorg server. Think of display servers as the skeletal beams that a desktop is bolted to.
Ubuntu 17.10 tested Wayland waters, but although 18.04 shied away from Wayland, the fact that 18.04 seems to have Gnome under control means the Ubuntu flagship desktop developers can turn their attention to Wayland, hopefully catalyzing its evolution.
Many saw the end of Unity not so much as an admission of defeat in cementing Ubuntu’s own desktop vision, but as evidence of a pivot in Canonical’s focus to cloud computing and IoT. After months in the wild and the update to Ubuntu’s incremental patch, 18.04.1, it is clear by this point that the decision to abandon Unity did not so much as jostle the stability of Ubuntu’s release. In fact, 18.04 has proven exceptionally stable, polished and well-received.
Few are the distributions that can put out as robust and distinct a product as Ubuntu, while also maintaining their own desktop. The only one that might lay claim to this is Linux Mint, but its code base has far fewer deviations from Ubuntu than Ubuntu’s has from Debian. Put another way, Mint’s code base is similar enough to Ubuntu’s (Mint’s upstream) that it can afford to dedicate time and resources to in-house desktops.
Without its own desktop, Ubuntu doesn’t seem worse for wear, but as refined and dependable as ever, especially with the introduction of features like a minimal install option and restart-less kernel updates.
It will be hard to tell how the end of Unity ultimately will impact Ubuntu until its next LTS drops in April 2020 — but for now, Ubuntu fans can breathe a sigh of relief as the distribution continues to shine.
Linux Gamers Won’t Be Steamed at Valve Much Longer
Another major development in desktop Linux computing was Steam Play’s August announcement of beta testing support for running Windows games on Linux. Steam evidently has been playing the long game (no pun intended) in backing work on the Windows compatibility program Wine, as well as the DirectX translation apparatus Vulkan, over the past couple of years.
This past summer, we saw these efforts coalesce. In a framework called “Proton,” Steam has bundled these two initiatives natively in the Steam Play client. This enables anyone running a Linux installation of Steam Play (who is enrolled in the beta test) to simply download and play a number of Windows games with no further configuration necessary.
A marked lack of access to top-tier games long has been a sticking point for Linux-curious Windows users considering a switch, so Steam’s ambitious embarkation on this project may prove to be the last encouragement this crowd needs to take the penguin plunge.
Steam has been exercising patience, as it has been maintaining a periodically updated list of the number and degree of Linux-compatible Windows games in its library of titles. It hasn’t been afraid to acknowledge that a number of Windows games still need work, another sign of sober expectations on the part of Valve.
Taken together, these steps suggest that Steam is in this for the long haul, rather than throwing together a quick fix to increase revenue from Linux-bound customers. If that weren’t proof enough, Steam even has gone so far as to post the code for Proton on GitHub, which is as good a sign as any that it is invested in the Linux community.
The entire undertaking holds promise to steadily improve the Linux desktop experience as more games reach mature compatibility, and Proton slowly crawls out of beta.
Red Hat Hangs Its Hat on IBM’s Rack
Although the Linux desktop landscape saw modest but undeniable progress, there was much more at play in the enterprise Linux arena.
Perhaps the single biggest Linux headline this year was IBM’s acquisition of Red Hat. IBM and Red Hat have enjoyed a long and fruitful partnership, and IBM’s shrewd tactic in competing with Microsoft more than a decade ago played the leading role in Red Hat’s rise in the first place.
Red Hat popularized, if not pioneered, the practice of selling support and tailored configuration as an open source business model. Fatefully for Red Hat, IBM was the big ticket customer that supercharged its revenue stream and confirmed the profitability of premium support. IBM minted its alliance with Red Hat because it wanted to compete with Microsoft in the server market without having to license an expensive operating system.
In some ways, IBM’s outright purchase of Red Hat may have been inevitable. The two have grown symbiotically for so long that subsuming Red Hat into IBM likely was the only way to squeeze more efficiency and return on investment out of the relationship.
You could even liken it to a couple who’ve been together for years finally announcing their engagement. Whatever else Red Hat’s purchase signifies, it legitimates Linux as an enterprise powerhouse, and lends credence to open source developers who long have touted the profitability of their work.
Amid all the deserved fanfare surrounding this betrothal, little attention has been paid to the reverberations it will send through the bedrock of the entire Linux space. Red Hat spearheads development of systemd, a replacement for the System V Linux init process that already has seen significant adoption among Linux distributions. This is no meager contribution, as the init system is the single most central component of the operating system after the kernel, and it dictates how the OS finishes booting.
Thus, the question on the minds of those who are giving this matter serious consideration is this: How will entrusting a (now) corporate-owned company to build the init process implemented in the vast majority of Linux distributions impact the course of Linux’s development?
Systemd of a Down
This leads perfectly into the next big story from the past year, because it demonstrates both the weight of the responsibility bestowed upon Red Hat in writing an industry standard init system, and the potential for harm, should this responsibility not be approached with proper humility and care.
Recently, a major bug affecting systemd was discovered. It allowed a user with a UID number higher than a certain value to execute arbitrary “systemctl” commands without authenticating, granting what amounted to full root access to that UID. The bug in question isn’t in systemd per se, but it pertains to systemd, in that systemd implicitly trusts the program containing the bug, polkit. So, because implicit trust itself is an unwise software development practice, to say the least, it equates to a bug in systemd, in some ways.
When systemd first took hold in the Linux biome, there was more than a little griping in the community. The central issue was that systemd contradicted the Unix philosophy by constructing and relying upon such a monolithic program (moreso than init intrinsically is).
To give a sense for how truly behemoth systemd is, it has swelled beyond the bounds of init’s reasonable purview to encompass DNS server IP assignment and regular task scheduling, relegating such venerable Unix stalwarts as /etc/resolv.conf and cron to (eventual) obsolescence. It seems that these Unix philosophers may have had a compelling, but ultimately unheeded, point.
Microsoft Opens the Open Source Patent Floodgates
IBM was not the only one to stake a claim to Linux: IBM’s perennial foe, Microsoft, made Linux maneuverings of its own in 2018. In October, Microsoft joined the Open Invention Network (OIN), subsequently open-sourcing more than 60,000 patented pieces of its software.
The OIN is a coalition of partners committed to insulating Linux and Linux-based projects from patent lawsuits. To that end, all members not only are obligated to openly offer patented software for public use, but also are allowed to freely license patents from one another.
Aside from the benefits this obviously confers on Microsoft, especially with companies like Google for fellow members, it puts another power player squarely in Linux’s corner. This may be the final sign of good faith the Linux community needed that Microsoft sincerely has embraced Linux and, moreover, that it has substantial plans for Linux-related projects in its future plans.
Open Source and Open Silicon?
There is one more notable milestone on the desktop Linux front — notable for what it portends for Linux, and computing on the whole. System76, the foremost Linux-focused hardware manufacturer in the U.S. (and maybe the world) has announced a line of high-end Linux desktops featuring open hardware specifications.
The Thelio line boasts an elegant, premium look that is sure to lure more than the privacy-conscious. Open hardware is the hardware analog to open source software, and while it has been an aim of the security-conscious and freedom-loving tech denizens, it has subsisted as little more than a pipe dream until recently.
The quest for open hardware arguably was accelerated by the Snowden disclosures, and the extent to which they revealed that hardware OEMs may not entirely deserve users’ trust.
Purism was the first consumer-oriented company to take up the charge but, as it will admit, its product is a work in progress, and not as open as the company and its privacy crusader allies envision.
Bringing more open hardware options to consumers, and thereby injecting competition into an otherwise sparse field, is an unalloyed good.
While reviews of the year’s events certainly are interesting, if just for a sense of scope, retrospectives aren’t particularly useful unless they are applied. With all of these 2018 milestones in mind, what trajectory do they suggest for 2019?
Last year easily was one of the best years for the Linux desktop sphere since I started using Linux (which admittedly wasn’t very long ago). Alongside big news from Steam and a reassuringly strong LTS release from Ubuntu, came piecemeal strides by distros like Elementary and Solus in solidifying their work and their reputations as just-works, mass-appeal desktop systems.
Along with the production of first-class hardware like System76’s Thelio PCs, and even Manjaro’s Bladebook, desktop Linux has never looked better.
I won’t indulge in the clich and predict that 2019 will be “the year of the Linux desktop,” but I foresee it building on the gains from 2018 to make even sleeker, more modern, and more usable desktops with burgeoning appeal outside the Linux niche. 2018 saw some high-profile publications giving Linux an open mind and a positive reception, so it wouldn’t be a far-fetched scenario for Linux to see an uptick in first-time users.
The enterprise realm is set to be much more tumultuous, as IBM and Microsoft have planted their respective flags in different corners of the Linux world. This could precipitate a wave of innovation in Linux as established corporate powers poise themselves for cloud supremacy.
On the other hand, this cloud computing contest could lead development of Linux and its satellite projects down a path that is increasingly dissonant — not just with Unix philosophy, but with the free software or open source ethos as well.