Recently, Linus Sebastian & Luke Lafreniere of Linus Tech Tips fame, endeavoured to replace their daily-driver Windows installs with their choice of Linux distribution. This was first proposed on their WAN Show podcast, and while it wasn’t discussed in depth beyond the basic concept, it generated a fair amount of buzz in the wider Linux community.

What is quite telling about this challenge, and when it was first proposed, is that some members of the commentariat were already pointing out that one of Linus’ favourite games - Anno 1800 - didn’t have a ‘Gold’ rating on ProtonDB, at the time; Adobe Creative Cloud and it’s lack of support was also brought up; as was the lack of proper HDR, in reference to potential out-of-the-box blockers to their regular workflow and experiences.

Given Linux’s historical and negligible market-share of the desktop space, it’s not overly surprising that a popular YouTube channel announcing a “Linux challenge” generated a good amount of interest, as well as worrywarts.

In this blog post, or series of posts depending on the extent of the drivel I end up spouting, I will be making a serious effort to look at their efforts with an objective eye, though with an admitted bias from the start, as I’m no fan of Windows and Linux is already my daily-driver, so take everything I say with a handful of salt.

There has also been a reasonable amount of discussion about the challenge on the WAN Show, in-between the main episodes being released to YouTube. It’s somewhat sporadic, but some information and comments are pertinent, so when I remember, I will also include links and references to these shows, though the bulk of the response will still be targeted towards the main ‘challenge’ videos.

Linux HATES Me - Daily Driver CHALLENGE Pt. 1 - 2021-11-09

Incendiary title aside, our first episode came out some time after the initial challenge had been discussed. Prior to it landing on YouTube, it was released on Floatplane as well as having been discussed on at least three episodes of the WAN Show. These small reveals, and titbits of information about what was in the first couple of episodes, left the community further intrigued as to the hosts’ experiences.

Several members of the community started to speculate on the problems that had been experienced already, because there were a few, as well as the potential solutions that they might have discovered. If not covered in the rundown of the main video, I will revisit any missed points at the end of this post.

Linus

Outside Help

Right from the start, Linus admits that he’s never made an attempt to daily-drive Linux, and that one of the reasons is the “paralysing number of choices you have to make, before you even get started.” He says that he could “call on industry contacts, or use internal resources” but won’t, because he wants to have the same experience anyone else might have.

I’ve gone back and forth on my feelings about this, because the truth of the matter is that a lot of people learn Linux, and other OSes, by being introduced to them by others, or university courses, or even through work. Linux, despite its small market-share, is still commonly learned by people who have resources like friends or family to help them out. That’s not the case for everybody of course, momentum has to start somewhere, but to immediately dismiss the seemingly-knowledgeable resources he does have at his disposal, is an interesting choice. While I don’t actively push Linux on anybody, in any flavour, I have friends who know that if they come to me for advice on trying Linux out, I will offer my opinion and support if they ask for it.

When we compare this idea to Windows and macOS, we should first consider that the majority of people will never install an OS in their lives. Most humans will simply buy the computer that they buy, have a selection of things set up for them out of the box, and in the vast majority of cases, this has been a Windows prebuilt for decades. As a result of Microsoft’s dominant market position, Apple have fought an uphill battle for years when it comes to switching users away from Windows, they still offer support pages on migrating from a Windows PC to a Mac, alongside articles on moving from Android to iOS for the same reason. There’s a fundamental paradigm shift for most people when they change operating systems, and Linux is no different in this respect.

Linus, through some of his comments during this challenge, has referred to himself as a power user, and I think this is where we start to see some of the cracks in his approach. Because he’s good with Windows, and none of us have any reason to doubt that fact, he appears to assume that his skills will be translatable to Linux on the desktop. While trying to sound non-confrontational or big-headed about it, there are people who might consider me to be a Linux power user, but I would never try to apply that same label to Windows, a system which I haven’t used in anger for many years; when faced with a Start menu, I would consider myself a Windows novice at best, and an outright hostile infiltrator at worst.

Linus vendors, such as Red Hat or Canonical, could do with prioritising more documentation on switching from Windows (or macOS) to Linux.

Where Linux primarily differs is Linus’ next point, which is the sheer amount of choice that Linux offers, if you decide that you want to install a ‘Linux desktop’.

Distro Choices

In line with Linus’ approach to installing Linux, he starts his challenge by looking up different articles on recommended Linux distributions, which is a reasonably valid approach given his self-imposed constraints, but he does complain about the inconsistency of these published listicles. This observation is somewhat fair, but also falls into the category of there being no ‘single company’ driving Linux on the desktop, entirely due to its distributed nature. “Linux” is really something of a misnomer, as the word itself can mean a variety of things, depending on the context of the conversation, and the person you’re in conversation with.

If you search “Linux” in a search engine, one of the top results should be the page: What is Linux? - Linux.com, on a website which is run by the Linux Foundation (the non-profit consortium, effectively responsible for promoting Linux.) This page is confusingly technical, but does use the term “Linux operating system” despite the consternation this phrasing can cause. It is a fairly daunting read to anybody who might stumble across it, when seeking more information on “this Linux thing,” and I would go so far as to say this page needs a serious rethink, as even the explanation of distributions is half way down the article, and comes after terms like ‘bootloader,’ ‘kernel,’ and ‘init system,’ just to scare people away.

An easy action item for the Linux Foundation, might be to make this messaging considerably clearer for new users.

Other pages exist alongside the listicles, such as the What is Linux? | Opensource.com page, and crucially, some of the more common Linux distributions which you might have already heard about from various conversations from people in the know. You might find the Ubuntu, Linux Mint, or Oracle Linux homepages presented to you, though none of these will do a good job of helping you to understand which one you should pick. Out of the three, Mint probably does the best job when it comes to terminology.

In short, Linus’ complaint is valid, and we as a community need to do a better job of raising awareness of Linux, generally in the form of better “what is” pages, and landing pages for the more common distributions. Sites like DistroWatch can also help, though again, it could use a modern coat of paint to better highlight popular choices for new users.

Linus raises the generally confusing nature of what makes each distribution unique, and this is also a perfectly-reasonable criticism of the situation; the truth of the matter is that most people could have a perfectly serviceable experience if they simply chose to install Ubuntu, but individuals in the community are far more likely to suggest their “favourite” distribution, over the one that might be right for a newcomer, while ignoring the flaws that their favourite distro might have.

As a collective, Linux users should accept one or two “noob-friendly” distributions for recommending to newcomers.

Briefly, Linus states that “unless [he] really wants to get into the weeds,” he’s better off sticking with something based on Arch, like Manjaro, or Garuda. He also says that if he were looking for a challenge, Manjaro’s “bleeding edge” approach “calls to him,” both of which are interesting statements, especially given the events that play out.

I’ve seen this, first hand, and honestly I hate it. There is no reason, zero, that you as a Linux user should be recommending any Arch, or Arch-based distribution to a newcomer. You are not Valve Software. You can talk to your friends about Arch, while helping them to install Ubuntu or Mint, but you should not labour them with the Arch family tree right out of the gate, for the simple reason that there’s a high likelihood they won’t be sticking with Linux for very long.

Arch itself, is fine. I like Arch, in years gone I have daily-driven it and been quite happy with the experience, but I am also painfully aware that it can be a challenge. Not only does it have a technical setup process, even with the recently included Arch installation script, no two Arch systems are alike, and while that can be an extremely powerful factor for experienced users who are looking to get the most out of customising and understanding their OS, it can also be horrible when problems are experienced, and you have to ask for help or debug the issue in a vacuum.

Arch derivatives are a problem. If Arch is tricky because it’s a technical sandbox, then the derivatives are a sandbox in which a castle has been constructed for you, and you’re expected to maybe add a turret here and there, but not mess with the foundations too much, because something you don’t fully understand will leave you in ruins.

With a recent default install of Manjaro, running updates through their package manager threw up a warning that several random packages were going to be replaced, most people familiar with the concept might have a glance over the list, and be happy that nothing that they can’t fix is going to break; new users will not think the same, and will be immediately put off by messages like this, or worse, they’ll blindly click “Cool, go for it,” and be surprised when their desktop disappears.

Clearly, I’m showing a bias here, but I stand by the statement that we shouldn’t be recommending non-established, or overly-technical distributions to brand-new users. Even if they don’t have any issues in the first few days, weeks, or months, eventually they’ll have some core program changed underneath them, through a seemingly innocuous update, and their slightly-delayed-rolling-release-distro will break.

Ultimately, Linus starts with Pop!_OS, and we both agree that Pop!_OS is a ridiculous name.

Pop!_OS

Linus starts to install Pop, based on the fact it was generally top of the listicles, and while making a casual mention of the fact his Windows installation has “bugged-the-crap-out” just prior to his switching of OSes.

Pop also has the added advantage that they bundle and offer an Nvidia image of their installation media, to get around the notoriously tricky Nvidia setup on Linux systems, AMD and Intel both have their graphics drivers embedded in the Linux kernel itself, as open source software, while Nvidia choose to keep their drivers proprietary and installed afresh, alongside each new kernel installation.

Briefly, this is sort of an important point, and one which the pair of hosts come back to throughout their time using Linux (as well as afterwards.) Linus and Luke aren’t shying away from the idea that Windows doesn’t have its own issues, and others have been quick to highlight similar ways in which Windows might bug a new user too; one notable, if exaggerated, example being this Twitter thread.

Linus doesn’t actually go so far as to remove Windows from his main machine, he simply swaps out the drive. You could argue that this shows a lack of commitment right from the start, but we’ll let this small transgression slide.

Linus is pleasantly surprised that his Thunderbolt dock works, out of the box. Several members of the Linux community were worried about this aspect of his setup, as well as other elements which were deemed to be “exotic,” and it’s nice when things “just work” as you might expect them to, despite preconceived notions that something is going to be difficult.

Immediately, Linus sees some audio issues, and realises that his audio has been configured to use the GoXLR multichannel device, which appears to be a pseudo-device associated with the GoXLR. Speaking from experience, I completely understand why this happens, and I can understand that someone who is used to the Windows audio subsystem might find the presentation of audio devices in Linux to be confusing.

Audio on Linux still isn’t great, and while desktop distros do their best, oddities still happen out of the box, especially with hardware that’s outside of the typical.

My headset, when I first got it, worked perfectly with Pulseaudio. It shows up as a few pseudo devices, and has some hardware features which make it useful for me, even though it’s far from a “standard” headset. I was honestly surprised that the experience was so good, and I was very happy, until I tried Pipewire for the first time and realised that the profiles of my headset were completely screwed up. Thankfully, this problem was fixed upstream in short order, and I was back to my regular usage, but it just goes to show that the landscape is still rough in some very core areas of the desktop Linux experience.

Linus installs Hardinfo to see if his “device drivers” are installed, and appears to acknowledge that Linux doesn’t really work like that, though it’s not clear if he’s referring to the driver situation, or the lack of a clear, GUI, device-information panel. I’d never used a GUI hardware information panel before I watched this video, and I only installed a couple as a result of it, to see if they were useful. I can see their utility, and acknowledge that just because we’ve been using lshw or similar, for years, there is a space for an out-of-the-box GUI panel, which displays information that the user might wish to be aware of.

Distributions could, if they wanted to be user-friendly, include more GUI tools where traditional CLI tools have generally taken preference.

Now, now we reach the point of the video that’s probably caused the most debate in the community. Linus hits an issue installing Steam, because of a dependency conflict with the package of Steam on his particular release of Pop. When he can’t install Steam through the GUI, due to a confusing text box error, he attempts to use the CLI instead, having apparently found an installation guide that tells him what to type. Linus removes his DE… by accident, after forcibly typing Yes, do as I say! on the command line. He calls this necessity “stupid” right before his DE is removed, and he’s dropped to a prompt.

Right.

Okay.

How did this happen?

Linus downloaded Pop from their site, using their official images, and installed it on his computer. What he wasn’t to know was that the image he’d downloaded was a build of Pop which just happened to have a dependency bug centred around the Steam installer, itself available in the official Pop Shop. Based on the desktop clock in Linus’ video, we can assume that he was installing Pop on October 6th, as his NTP setup was probably working.

It appears that the issue with Steam dependencies causing an uninstallation of the Pop desktop was a known problem, at least on the 30th of September, though it doesn’t look like it was fixed at this point. Remember that the Linux challenge itself, was first publicly proposed on the 24th of September. After the video was posted on Floatplane, a Pop developer raised an issue that the error messages used were unhelpful to the average user.

There are two things wrong with this situation though, the first being that it’s unreasonable for an end-user to expect, or know, that the ISO they’ve just downloaded from the official sources, has a crippling bug in it, especially if that bug has been fixed upstream (though that doesn’t seem to have been the case here, at the time of installation).

The second issue centres around the general notion that Linus should have updated his computer immediately after installing. The act of an initial update should either be enforced by the distribution when it’s first installed, with workaround for users who “know what they’re doing,” or it should be heavily hinted at during the setup process. Anecdotally, users like to get right into the meat of setting up their system, after a fresh install, and might only go to install updates when they’ve done everything else first, including installing their game store of choice.

Have a look at the Reddit thread above if you want to see some of the fallout, and other people have commented on this situation at length, so I’d only be repeating what they said. Needless to say, changes have been made as a result of this fracas, and more changes will likely come in the future to “LSproof” systems.

I do not think that this was initially approached correctly by the Linux community, and some of the Pop representatives; but I also think that the targeting of certain developers exacerbated the problem. Collectively, we should acknowledge the mistake for what it was, on the part of Pop, and we should also acknowledge that mistakes do happen, while working to ensure they’re unlikely to happen again in the future.

Linus, regardless of what you think about him personally, wasn’t entirely to blame for what happened to him. He was a gamer, installing Steam, and that act removed his entire desktop environment. Pop, regardless of what you think about Ubuntu derivatives, or System76, also weren’t entirely to blame.

Back to the video, Linus tries a hard reset at this point, but gives up on Pop after he realises it’s a wider-problem than just him. It sounds like he’s done a bit of searching and found other people with the same problem, potentially with ‘steam,’ ‘pop,’ and ‘borked’ keywords.

Linus states that “the things [he] did were not entirely ridiculous or unreasonable” and he’s right.

Manjaro

I’m confused as to why Linus goes for Manjaro, but it’s probably because it’s recommended a lot by the community as a whole. See my earlier rant about Arch-based distributions and our inability to recommend sensible defaults to newcomers. Again, Linus didn’t have any issues with Manjaro personally, I just don’t think it’s a good starting point.

Linus has questions around the name “X Server” which is perfectly reasonable, because it’s a confusing name and concept. Even seasoned Linux users get confused by window managers, desktop environments, display servers, Xorg, X11, Wayland, KWin, Mutter, etc. etc.

Linus starts Steam through the built-in Manjaro install, he generally has a much easier time of it as Manjaro includes the kitchen-sink with its installation.

Linus goes to install Cave Story+. This is really unfortunate as the native version of Cave Story+ is both old, and not well-supported on modern systems. It opens incorrectly on his install, with the game taking up half of his display, and being cut off. Looking at the Steam store page for Cave Story+, it appears to be from 2011, and with a “Recommended” OS of Ubuntu 10.04, which as you might have guessed from the name, is over a decade old.

ProtonDB recommends that users simply play the Windows version of the game, using Proton, but Linus wasn’t to know that, and a new user can’t be expected to know that either (especially when a native build of the game is available).

Valve and developers could possibly do more to ensure that Native Linux games continue to work, beyond their initial build and release.

Cave Story+ also seems to have issues with his attached Xbox One controller.

Linus moves on to Valheim, which is also native, but much newer, and doesn’t appear to have the same issues as Cave Story+. He comments that his Xbox controller also works, but that sound still isn’t.

At the end of the video Linus has realised what he did to break Pop, but he does refer to the error messages that he received as jargon which the average user wouldn’t understand.

Luke

Linux Mint

Unlike Linus, and because of prior-knowledge, Luke goes for Linux Mint, he’s used it before and daily-drove Ubuntu for a couple of years in school, so he’s comfortable enough with Linux in general that he’s confident in his choice. It would be unreasonable to compare Luke and Linus’ experiences directly, as they’re clearly at different starting points, but it’s still interesting to look over Luke’s experiences.

The video itself is mostly Linus, for obvious reasons, and Luke doesn’t get much of a look-in, honestly he doesn’t seem to have many problems, though he still has some. He is a user who “work[s] in development every day,” so it’s easy to say he might be a little more qualified.

He suggests that he’s going for Mint because Linux chads (nerds) terrify him, and he wants the challenge to go well, because he still wants the viewers to like him at the end of the experiment. It’s reasonable to go with what you know, and Linux Mint is about as friendly of a distro as you could choose to start with. It has a pretty reasonable base in Ubuntu (unless you go for the Debian edition of Mint, which Luke didn’t,) and it’s fairly inoffensive to most members of the community.

Installation

Luke does come across an interesting issue in the Mint installer, where he appears to have a single display (X?) across both of his physical displays, and it’s misaligned. He ‘fixes’ this by turning off one of his monitors and starting again. This seems to be one of those small bugs that’s tricky to fix, though trivial to work around. One thing that the Linux community as a whole is pretty bad at, is chasing features instead of squashing bugs.

You could easily make the argument that it’s a lack of drive to fix bugs that means they don’t get resolved, and it’s been suggested that because development is mostly done “for free” in an individuals “own time” it’s understandable that fun stuff gets done, instead of bug squashing. Of course, Red Hat still has bugs that go back years, and that’s funded by one of the biggest tech companies in the world, so I’m not sure the argument has much weight, I think developers just like to write new stuff, not fix small, niggling issues. I get that.

He’s happy with Mint’s user-friendliness, you get the option to install additional codecs on install, for example. There’s a whole post to be written around the concept of codecs and distribution, but it’s not a conversation for now. The fact that it’s a checkbox in the Mint installation is nice.

Post-install

Luke changes from the Nouveau display drivers the official, proprietary Nvidia drivers, reboots his machine, and it works.

Luke runs updates as his next step, which is amusing given my previous comments. This could be argued that it’s because he’s used to the concept of installing, then updating as a first-ish priority on Linux distros.

Mint places his displays the wrong way round, meaning his monitors are backwards. This issue is trivial and not exactly something that’s easy to solve, as your computer has no concept of where you’ve physically put your monitors. This is a ‘set and forget’ step, once.

Luke immediate goes to the Lutris website (after realising that it’s not available in the Mint software manager,) and installs Lutris by following their guide. It’s interesting that he knows to do this, and I have to assume that he’s simply recalling one of LTT’s previous videos and knows that it’s a reasonable tool to install if you’re going to be doing a lot of gaming on Linux, or he’s been on enough of the Linux community boards to understand the same.

Luke does have a problem with the software installer where the install of Steam doesn’t seem to complete, even though it actually has in the background.

Luke logs into Steam, enables Steam Play (Proton) for all other titles in his library.

Luke installs FTL, the native version, and then starts it.

Rounding out his segment, Luke says that he’s done the first challenge, and it really wasn’t a challenge, which is amusing when compared to the amount of issues that Linus had.

Elsewhere, Luke talks about a bug that he did encounter with his windows lagging on the Cinnamon desktop. It turns out that this issue is longstanding, and simply hasn’t been addressed.

Conclusion

The bulk of this breakdown obviously focused on Linus, and his own problems when trying to perform the simple task of installing a Linux distribution, and playing a game. For better or worse, this video series has, and will continue, to shine a light on the Linux community, as well as Linux as a viable operating system for both day to day use, and gaming. Already we’ve seen interest and posts from prospective users on community boards, entirely driven by the LTT videos, and there continues to be a steady uptick in users reported by the Steam Survey, some of which may be as a direct result of LTT too, though the Steam Deck definitely plays its own part.

Obviously I’m happy that changes have already come about because of this series, though I’m wary in the same breath; I believe that more users will be a good thing for overall attention from companies, in both the hardware and software spaces, but I also know that more users means more strain on developers, especially if those users are overly demanding, and without some financial incentive it can be very difficult to drive yourself to work on a project, when people on the other side of the planet are aggressively capslocking at you.