While it’s true Steam Machines aren’t exactly flying off the shelves, our reasons for striving towards a competitive and open gaming platform haven’t significantly changed. We’re still working hard on making Linux operating systems a great place for gaming and applications. We think it will ultimately result in a better experience for developers and customers alike, including those not on Steam.
Through the Steam Machine initiative, we’ve learned quite a bit about the state of the Linux ecosystem for real-world game developers out there. We’ve taken a lot of feedback and have been heads-down on addressing the shortcomings we observed. We think an important part of that effort is our ongoing investment in making Vulkan a competitive and well-supported graphics API, as well as making sure it has first-class support on Linux platforms.
Valve has done a lot for Linux gaming, and it’s good to hear they pledge to continue doing so.
For those curious on why Valve posted this, it’s because I wrote this article:
https://www.gamingonlinux.com/articles/valve-has-removed-the-steam-m…
Then all the big sites picked it up.
Valve’s support of Linux gaming has been almost flawless. Really they’ve done a fantastic job. Getting Sam Lantinga from Blizzard/Loki Games and focusing on the drivers and support infrastructure has definitely been the right move.
The biggest barrier to Open Source Linux gaming at the moment remains modding/developer tools for artists. There’s a lot of game engines and a lot of ported titles, but workflow issues remain e.g hardpoint editors, mission editors etc. This is what’s stopping open source game development from taking off in a big way. The example I like to use is Freespace 2, a 20 year old open source game engine which only has editing tools available on Windows. The Mission editor FRED2 only runs on Windows and while the asset import tool for ships now runs on Linux, the editors and viewers for the game’s archive files are still Windows only. Descent 1/2 and 3 all have level/asset editors only available on Windows despite these tools being open sourced. This needs to change before modders will jump in and create new content. Modding is a natural fit for open source game development, get the modders and you’ll see a lot more games and content. Modders are future game developers.
The commercial industry moving to standard engines has been a massive boon for the platform, however there is still a lack of AAA titles. I think there’s still room for another porting house similar to Feral/Aspyr but focusing on different titles. Linux gamers have a lot of choice in Indie titles, but are still lacking in choice in non FPS/Racing/RTS genres. A solid AAA RPG is still lacking on the platform of the likes of Oblivion/Fallout 4.
Edited 2018-04-04 22:46 UTC
Linux will continue to lack a library of AAA games for several reason, one of which being the lack of enough market to bother. And there’s nothing to debate in that regard because all aspects of reality have been speaking loud & clear year after year after year after year after year after ……..
ilovebeer,
You are right about the niche market status, clearly AAA games will target the largest markets. However I think this argument overlooks a detail that game developers don’t have to target linux specifically to support linux. They can get linux for free by targeting a multiplatform game engine to begin with. They could realistically target linux and windows in one go.
I’ll admit that I don’t have game industry experience, but if I were going to write a game it seems to make sense to me to use a multi-platform engine rather than a windows specific one.
Almost none of the engines are platform specific.
The fact is that the player base is tiny compared to the rest of the market and it simply isn’t worth the bother supporting Linux and even then if they do it is going to be Ubuntu. It is simply a cost benefit analysis whether it is worth supporting the game after development.
https://www.gamingonlinux.com/articles/linux-game-sales-statistics-f…
Some claim that upto 25% of their support time is spent on Linux alone, when it is less than 5% of their player base. With stats like that I wouldn’t bother.
The only reason a lot of game engines became cross platform was because of Android and iOS devices being popular.
Edited 2018-04-05 07:07 UTC
coherence,
Yep, there’s no doubt that popularity increases support.
The 25% figure is in article I linked, you obviously didn’t read it.
coherence,
Edited 2018-04-07 17:38 UTC
I think all major game engines, like Unity, support Linux exports so it should be zero overhead. I have tried that myself and it just works.
Edited 2018-04-05 06:24 UTC
Just having game engine support isn’t quite enough if you’re doing anything beyond the most basic games.
At a minimum, even if you don’t need anything platform-wise beyond the base game engine, you still have to have a way to build for Linux (not all cross-platform game engines support cross-building), and you still have to do testing on a Linux system. In reality though, very few games I’ve seen that weren’t purely independent ventures or done by very small companies have used just the game engine. Quite often, they end up tying in extra stuff designed to run on top of the game engine and provide bonus features, and a lot of that does not consistently operate on Linux.
Now, assuming that none of that is an issue, you’ve still got customer support overhead. Just because Linux users tend to be more tech savvy doesn’t mean they can solve all the problems themselves, and some really aren’t any more technically savvy than your average Windows or Mac user, so you need to account for not two but three platforms in your customer support flow-chart, which further increases training costs.
ahferroin7,
I think it’s a fair point, hypothetically. However do you have actual evidence that linux users need more support than windows users in practice? Ie if they are 2-3% of your users, do you have stats showing that they consume more than 2-3% of the project’s support resources? I’m genuinely curious if you have links with numbers.
You could probably mitigate most of the support requirements by going through a distribution service like steam, which takes care of the platform specific installation for you.
Doing this would of course solve the installation aspect, but that’s only a small part of support for a platform, and it’s one of the easiest parts to solve (even without using a dedicated distribution service). Keep in mind that for any reasonable developer, saying ‘I support X on Y’ doesn’t just mean that X runs on Y, it also means that you provide support to help make sure it runs on Y, including handling bug reports (which becomes exponentially more interesting the more platforms you support).
ahferroin7,
What you say is true, but it’s also true that in most cases someone else has already done the grunt work. So I guess it depends how much of the wheel the game devs want to reinvent versus re-implement by themselves, haha. I’ll admit that I sometimes build my own libraries at the cost of more debugging.
self,
Oops…meant reinvent versus reuse, but somehow I wrote that instead. Anyone know of a good debugger for our brain?
The library of “AAA games” for Linux is actually pretty good nowadays, in good part because of Steam. GoG and Humble Bundle helped too, but not so much for AAA. For what it’s worth, the MacOS offer is about the same.
I’m a regular player and haven’t felt the need for wine for the last few years (and haven’t used Windows for over 17 years).
There’s always going to be some titles missing or arriving after the hype has died down, but there’s already more quality titles available than you have time to play. Some will find the unavailability of $FAVORITE_GAME a showstopper, but most people don’t care as much and would be pretty happy with today’s Linux offering.
As both a Linux user and a gamer, you should know better than to claim Linux has a `pretty good` library of AAA games. I don’t know why you would intentionally walk into that quicksand but there’s no question that’s what the claim is. As far as MacOS – who cares? Steam Machines specifically target Linux, and that’s the subject here, not MacOS.
This is really not that complicated. Linux gaming has been a mess since forever for a myriad of reasons. It’s noble of Valve to try to improve it, and most of us can appreciate their efforts, but at the end of the day deep down inside most people know those efforts will never transform Linux into a formidable gaming platform. For people who want to argue in favor of Linux gaming, I say simply that the numbers don’t lie. Linux can’t be both great for gaming with a good AAA library, and a flash in the gaming pan. The numbers prove the latter.
Another point… If using multi-platform game engines and cross-compiling makes it so easy & cheap to get games to Linux, why aren’t the huge developers taking advantage of the easy money? It takes more than creating working binaries to get the performance needed. There’s still a lot of development that takes place outside of the game engine in many cases, and it can be a significant amount – or to the game engine itself so it provides what the developers want from it. Game support and customer support are two other huge areas we haven’t even gotten to.
Linux is not the gaming gold rush waiting to happen so people have convinced themselves it is. It’s not even a good money-grab. It’s more like collect cans and bottles. Technically you can make money doing it but it’s going to be a lot of work and in the end there’s very little chance it’s going to make you rich.
As Valve just admitted, after 5+ years of heavy investment, “Steam Machines aren’t exactly flying off the shelves”, and they were supposed to be on par & real competition for consoles and Windows.
ilovebeer,
While I disagree with you that it’s much more work for development, I agree with you that the market for paying linux users may be relatively weak. IMHO there’s always been this conundrum for developers to target linux.
It’s somewhat like android vs IOS. For reasons that have much more to do with marketing than with the underlying technology, android’s lower price points attracts poorer users on average. The IOS market is smaller, but it’s users have already demonstrated a willingness to spend lots of money on their gadgets. Android still gets lots of support due to it’s dominant market share, however linux desktops don’t have this benefit. Not only is the linux market a small niche segment, it is also comprised largely of FOSS users who have a reputation for not wanting to pay for software.
Edited 2018-04-05 16:08 UTC
Those are all good points that shine yet more light on the problems. To expand further regarding paying customers, micro-transactions have become commonplace in the gaming business model. So, if much of the Linux user pool isn’t keen on paying a one-time fee for a game, how in the world do you get them to pay repeated small fees, which is even less attractive? For companies who survive off micro-transactions, that creates a serious obstacle and easily puts their products into a high-risk category on Linux. From a business standpoint, the choice is fairly simple – do you want to fight hard wars for crumbs at best, or not bother and just go where the money is?
Trying to shoehorn an industry onto a platform where they’re at fundamental odds with each other is well beyond `an uphill battle`. I’d like to be optimistic on this subject but there just isn’t any compelling, much less convincing, argument I’ve heard yet.
Sounds like you’re in the “through of disillusionment” of the hype cycle. I feel Linux for games is more towards the plateau of productivity. It doesn’t have to overtake Windows in order to be successful. It doesn’t have to be sucessful in order to be good.
I can’t speak for game developers (appart from reading various blogs that seem to point out that supporting Linux is extra work but isn’t *that* hard), but as a regular gamer, I feel that Linux is a “good enough” platform today.
Before you start explaining to me why my feeling is incorrect, let me define “good enough” as “I don’t feel any need to look for a better one”. I have plenty of non-gaming reasons that drive me away from Windows and towards Linux, and the gaming reasons that would drive me towrds Windows are too weak.
So, game developers, please make more Linux-compatible games. I’m not likely to move to a different platform. I have stopped buying games that I can only play through wine, but I do buy Linux-native games every few months.
Why would I say your feelings are incorrect? In fact, I just stated the opposite. Again, the conversation isn’t about you, your feelings, or if Linux gaming suits your needs. The conversation is much bigger than that. We’re talking about this at an industry level.
The conversation was about the size of that library, and by extension how suitable is Linux as a gaming platform, today. And even if you claim that it is “bigger than that” and “at an industry level”, in the end the assessment is still subjective.
The problem is that all those Linux games in that list are mostly older than 3 years.
The last major AAA game that was released on Linux that I can rememer is Deux Ex: Mankind Divided, nothing since.
I agree, there is nothing wrong with Linux and gaming, in fact I normally game only on Linux. It would be in the interest of Steam/Game publishers to release more for Linux as I keep playing Borderlands 2 now.
No it isn’t. Market share is not subjective. Sales figures aren’t subjective. Cost-benefit analysis is not subjective. A companies viability to remain in business isn’t subjective. Regarding gaming, Linux is minuscule, is very poorly supported, and has an extremely small user pool. Linux could vanish tomorrow and the video game industry wouldn’t even notice. The conversation is about why that is and a game library is only one factor of several.
[q]We are talking about the lack of AAA titles on Linux, not filtered lists you create on Steam, or games you find entertaining. It’s beyond laughable you think 1/3 of the most popular games are available to Linux.
Take a list of the top 20 most popular game franchises, or even the top 100 best selling games, tally which are available for Linux, and be ready for disappointment.[\q]
That’s precisely what I did: The list I took is the top 100 best-selling games on Steam, not my own curated list of favorites. And it has 33% Linux-supporting games. If you think my methodology is wrong (it’s clearly not perfect), you’ll need to provide a better one. Or is your mind tricking you into ignoring facts that contradict your worldview ? That would be disheartening but common, which I guess fits the “beyond laughable” description.
There doesn’t seem to be much point in giving a detailed response to the rest, but one last tidbit: you’re conflating Linux gaming’s market share (an objectively tiny one indeed) with the suitability of the platform (a more subjective value, I find it pretty good despite its various issues including available titles). Finding it pretty bad would be reasonable too, but I find your have an irrationally negative view of it.
My view is the view shared by the vast majority of gamers, and developers for that matter, so you must think we all have an “irrationally negative view”. It’s hard to take comments like that seriously being as absurd as they are.
I’m not surprised you’re not up for the challenge when I said, “By all means, do your best to make a good argument *against* going with platforms that 1) are vastly popular, 2) have a vastly larger library of games available, and 3) in the case of consoles, are cheaper and require NO maintenance.” It’s a fools challenge in that only a fool would actually think there *is* a good argument against it.
Lastly, no, I’m not conflating Linux desktop market share and Linux suitability as a gaming plaform. They are obviously two different things, but obviously both factors in the big picture. Linux desktop market share is tiny, therefore the potential customer pool is tiny, therefore the profit potential is tiny. The Linux desktop itself along with various underlying components and subsystems aren’t in or don’t stay in a state suitable for a stable serious gaming platform, therefore you have increased cost in both development and support. Making a Pros & Cons table it quickly becomes easy to see why Linux gaming is in the state it’s in. And with that I’ll remind you that being less horrible than it was 10 years ago doesn’t mean it’s good today. But hey, you know, this is finally the year of the Linux desktop so it may as well finally be the year of Linux gaming while we’re at it, right?
Come on now, Linux simply hasn’t got a lot of AAA titles. While it has gotten better by leaps and bounds in the last 10 years there’s still not all that many AAA titles.
I have 100 games on Steam and out of those 34 runs on Linux. Of those 34 the only ones I’d consider AAA titles are Left4Dead, Bioshock Infinite, Tomb Raider and DiRT Showdown. That’s 4 our of 100. If we also count Serious Sam Fusion, Metro 2033 and Portal 1 and 2 there’s a total of 8. That’s not a lot.
Of course, most of these titles are unplayable in Linux for me anyway because if you don’t have an NVidia card gaming in Linux is ass.
The only reason I still dual boot to Windows is to play games and I’m pretty sure I’m far from alone.
Edited 2018-04-07 05:55 UTC
I fully agree that it lacks a lot of titles, and that this’ll be a showstopper to some.
But to me the current offering still amounts to “pretty good” in the sense that I don’t feel the need to look elsewhere (wine, dual-boot…) anymore since a few years. I get more than I need for my gaming fix, and I know that others (not everyone) can be / are just as satisfied.
“…plus you’ll get other advantages by using a Free system”
Except the closed source Nvidia drivers.
If the editors have been open sourced, what prevent them now from being properly ported to Linux ? Lack of motivation and/or talent ?
Package management.
That was true a couple of years ago, but now we have (more or less) universal packages like Snappy, AppImage, and Flatpak. They are WAY easier to create than some PITA Debian packaging.
Edited 2018-04-05 06:29 UTC
The fact that you named many different systems, and that I haven’t heard of them, kind of proves my point.
They’ve just replaced one package management problem with another.
Which would a company choose, and can the company be assured that it’s not going to run out of steam (hah) and get abandoned? Will those package management systems interfere with each other? Will the company have to support all of them?
It kinda doesn’t though since there are many different systems for creating installers in Windows (InstallShield, WiX, NSIS etc etc etc holy shit there’s a lot of them) which is basically the same as the package management “problem” in Linux.
Yeah, but the point is by now, most of them work without error and don’t step on each other’s toes. The only installer for Windows that has given me any trouble in the past few years was, ironically, IBM’s piece-of-shit Installation Manager for installing Java* programs.
* Not just any old Java programs, but their Eclipse-based Rational development environments. Eclipse is already bad, but you can always count on IBM to manage to outdo an existing piece-of-shit with something even worse.
The universal packages are not “stepping on each other’s toes”. You clearly have no idea what you’re talking about. You might get into trouble if you install 5 different Firefox packages.
Edited 2018-04-05 10:46 UTC
And the package managers doesn’t step on each others toes either and most of them work without errors.
package management is a non-issue. The reason for lack of editors and other mod tools in Linux is because it’s not a walk in the park to port Windows applications that wasn’t written with portability in mind to GTK or QT.
Why they aren’t “simply” compiled against winelib?
If you haven’t heard of Flatpack, Snap etc. then you obviously don’t follow Linux or related news very much! It’s been quite huge news for a couple of year now and many articles have been published on every Linux news site I have come across.
Basically, they are the Portable Apps for Linux, you just download the program, double click it and run it, no installation required.
More importantly, no package manager is required either, which makes your point rather moot.
No it doesn’t. I was talking mostly from the point of view of the company that makes the software. Which option do they go with, and how can they get assurance that the one they chose will still be supported?
It doesn’t matter which option they pick. That’s the point of universal packages. And if the option becomes obsolete then they can take the one that is relevant. The packaging itself is not rocket science (like Debian packaging).
Besides, for example Google Chrome is available only as a DEB and RPM and I bet that covers 90% of Linux users.
Edited 2018-04-05 10:39 UTC
> Which option do they go with, and how can they get assurance that the one they chose will still be supported?
Flatpak. Ignore anything Canonical invents in house because it all gets abandoned sooner or later. Flatpak is the only non-Canonical one that does meaningful sandboxing and has an isolation layer between the OS and the applications. Appimages are literally just iso images like mac DMGs, so you have all the same problems of no proper package management of them, whereas with Flatpak you can set up a repo and push updates through that, rather than inside the game itself.
Flatpak also has a promising pedigree, beginning as xdg-app from Freedesktop,org… the same collaborative effort between KDE/GNOME/etc. developers that produced such successes as:
1. xdg-utils (The utilities like xdg-open and xdg-desktop-* which things like the GOG installer rely on to Just Workâ„¢.)
2. D-Bus (Replacing KDE’s DCOP and GNOME’s CORBA-based equivalent)
3. The XDG Icon Naming, Desktop Entry, and Desktop Menu specifications, that allow all major desktops to share icon themes and desktop entry definitions. (Thank goodness. I still remember what a hassle it was when KDE 3 and GNOME 2 had their own incompatible sets of icon names and launcher registries.)
4. xdg-user-dirs (The desktop-agnostic configuration mechanism for specifying where applications should look for Linux’s equivalents to things like “My Documents”)
In this case, isnt steam itself kind of the solution to this for valve?
Package management hasn’t been a problem for years.
Want your game on Linux? Steam client does it for you, GOG package Linux games themselves with MojoSetup, itch.io also has a client like Steam.
For normal packages outside of those stores, AppImage has been around since 2004 – Beamdog (reviving lots of old RPGs like Neverwinter Nights) use AppImage on Linux for their own client and it works perfectly. Snaps and Flatpaks are newer, but also work across different distributions.
Packing is not an issue, anyone saying it is just isn’t looking. Anyone claiming it still is despite them, obviously glossed over the many ways to differently package for windows.
Edited 2018-04-05 09:36 UTC
From what I’ve seen/heard it’s mostly a lack of talent, lots of good Linux programmers, but they have no interest in porting MFC applications to GTK/QT. On the other end the people who want to use the tools can’t program (artists) so they have to stick to Windows to get work done. Package management has nothing to do with modding tools getting ported. It also has nothing to do with games being ported/released. Self extracting installers solved that issue years ago. Loki Games solved that in 1999.
Edited 2018-04-05 08:48 UTC
The only thing i care about Steam and Linux is a good replacement to Moonlight, so i can buy graphic cards that are not Nvidia
http://steamcommunity.com/app/353380/discussions/0/1696043263494088…
Linux is the Valve’s safety… valve. In case of MS or Apple trying to limit software distribution on their respective platforms, Linux Steam machines are an option.
I just bought Tomb Raider (2013) and Painkiller: Hell & Damnation (2012) through Steam on my Arch Linux computer this week. They both run great!
I’ve been a Linux user since the time when “Linux gaming” meant “Tux Racer and Nethack”, so the fact that I can play these games on Linux still blows my mind.
Same here, I’ve recently bought Tomb Raider and Life is Strange, neither of which were available on Linux originaly.
I don’t care that I didn’t play them when they came out:
* I generally wait for the price to go down anyway
* these games didn’t age much, are just as enjoyable today
* my game library is already much bigger than I have time for, I’m never short of a great game to play
Life is strange was good, i really enjoyed, and that caught me by surprise. Hope you do to
Galbraith’s Law of Politics: Anyone who says he isn’t going to resign, four times, definitely will.