Torvalds: SteamOS will "really help" Linux on desktop | Hacker News Reader

Torvalds: SteamOS will "really help" Linux on desktop

pcpro.co.uk

The reason Linux hasn't seen widespread adoption isn't just that it doesn't have the marketing power of Microsoft behind it. It's flaky, it lacks support for a wide variety of hardware, it requires a level of understanding that the vast majority of the population just don't have time or willingness to gain, it lacks obvious usability features in the most recent versions of major distros.

The general population requires their computer to make their life easier, Linux just doesn't do that. I've been hacking around with Ubuntu now for the past month and I'm still not satisfied with it. I have a laundry list of things that are far from satisfactory that I don't yet have the capability of fixing myself... and I've been developing software for 20 years. If the learning curve is this steep for a long time developer, imagine how daunting it is to the general public. They don't get a computer because they want something to tinker with and figure out how it works for hundreds of hours a week, they want a computer to do something for them that will enrich or simplify their life in some fashion.

Perhaps if some of the more fundamental features were fixed instead of adding nice to have trinkets, you'd find more people will take it more seriously. I shouldn't have to go and hack around with complicated and potentially not-yet-existing files, referencing incomplete or inaccurate blog posts written by the last guy who tried what you're attempting and got so pissed off with it that he thought he'd save the next idiot the headache he'd just gone through - only to omit the most important piece of information that's not to be found anywhere else on the internet.

That is why Windows has a major market. Because unlike Linux, shit just works... for the most part.

They say that Linux is free as in "freedom", but honestly, how free are you when you have to spend all your time tied to the computer just to get shit working?

I'm trying to love Linux, I really am, I'm on my 4th distro in 2 months and so far I feel like the only thing I've achieved is getting Stockholm Syndrome.

If Steam can fix this, I say please, please bring it on. Torvalds, I think you've got the right idea, honest to God I do, and I'm trying to love your system but it seriously needs some love from developers and usability experts that understand people.

Same here! I'm a technologically educated person, I use command line Linux for all my VMs and servers, I love it...but nothing "just works" when you jump to a UI powered by Linux.

So, I run a Samba share on my VMs and configure my desktop to be friendly.

Exactly, nothing just works... it's a giant f ball ache.

I know that once I get past the learning curve, it'll be okay; but my inner Jony Ive screams in agony every time something doesn't just work.

I must've spent 50 hours on my machine just trying to get LUKS to work properly so that my machine will suspend to disk so that when I shut the lid, it asks for my LUKS password again before letting me in, and it still doesn't work. When I lock the machine, my second monitor doesn't lock so anyone can still see what I was looking at before I locked the machine. It's agony every time I want to do anything more complicated edit my photos or sit on Hacker News.

But I really want to ditch Windows...

Android has millions of casual consumer users, and the state of that ecosystem is unmitigated shit. Worse free driver support there than the traditional distro scene for desktop computers had a decade ago.

Striving for widespread desktop adoption is a mistake. I see no reason to think that an influx of clueless users would improve anything at all. Selling steamboxes to the 'unwashed masses' may incidentally improve the situation for users desktop linux, but getting more users onto desktop linux should not be our goal or our strategy.

It very much depends on what the goal is. GNU/Linux is already extremely popular on server, on embedded devices and on smartphones in the form of Android.

Now why wouldn't we want it to become a desktop platform with mass appeal? It certainly would improve the situation for existing Linux desktop users insofar as commercial attractiveness is directly proportional to the size of the user base. A greater number of users could for instance make it more likely for game developers to port their products to Linux. It also forces hardware manufacturers to pay more attention to their Linux drivers or risk being outdone by their competitors. It could, furthermore, prompt companies to port their specialised commercial software (e.g. Photoshop) to Linux. All in all I'm unable to see the downside of strengthening the Linux ecosystem.

Linux on the Desktop is like owning your own Cessna. It's not something that you do unless you have a very good reason to. You need it for work (crop dusting)? Great, it can accommodate you. You need it because you have an exceptional use-case (live in Alaska)? Great, with some effort it can probably be adapted to work there too. You need it because you have an unquenchable thirst for power and control, or because you yearn for freedom? Yeah, it'll do that too. Do you need it because 20 miles is too far a commute to walk every day? What the hell are you thinking? Just get a car; the advantages of a Cessna are of no use to you, it can only be a hassle.

The flying car is a fool's errand. For people who need airplanes, or want airplanes, airplanes are still here. For people who just need a car, there is no good reason to make that car fly. If my mother asks me what kind of car she should buy, I don't tell her to buy an airplane. If she asks me what sort of computer to use, I don't tell her to run Linux.

The downside to a bunch of suburbanites commuting with Cessna's should be obvious. The downsides to a bunch of "suburbanites" trying to casually use Linux are admittedly less obvious, but still present if you think about it. How much money do you think Microsoft and Apple spend on customer support? The community sure isn't going to take up that slack, and I see no reason to think that companies like Canonical are up to that task (they can't even handle their current (skewed technical) user-base (see: other commentators in this thread). In practice, unskilled users put additional pressure on volunteer developers. Non-technical users are not free; they come at a price.

I love Linux on the desktop, I've been using it as my primary OS for more than a decade now. The flying-car/'Year of Linux on the Desktop' dream though? I've been there, done that, got the t-shirt, and turned the t-shirt into a shoprag.

there's no GNU on Android, and probably not on embedded devices either (given the GPLv3)

so you'd be more accurate to say just Linux, rather than parroting stallman

Correct. I like the phrase "GNU/Linux", but only in recent years with the popularization of other userlands that makes the distinction has become useful.

glibc is under GPLv2 and I find it to be the defining piece of GNU/Linux system. The coreutils are under GPLv3 and therefore projects like Tizen use pre-GPLv3 version of such software. So parts of GNU ecosystem are popular on embedded and the others are not.

You mean the Linux Desktop, not Linux. Torvalds heads the Linux kernel - it is already the most successful piece of software in existence with the highest number of users and very good user experience. Android is Linux, most routers are Linux, etc.

The Linux Desktop has nothing to do with Torvalds directly and needs, as you say, a dose of "just works". That "just works" isn't a technical thing as most people seem to believe though. "just works" is more about downloading an app for your airline and having the app do everything for you, and phoning the airline's help desk and having the help desk actually understand what you're using. This takes massive buy-in from society into a standard platform and a standard platform is something the Linux desktop has generally fought against tooth and nail for various reasons.

If Steam can provide that platform, Microsoft is going to be in serious trouble. It seems doubtful at this point though as Steam seems to be heading more in the direction of a console or Android-like implementation rather than a general purpose computing platform.

Okay, let me back-track a little and clarify my point. Yes, I mean Linux desktop. Linux server/routers/other operational devices has seen widespread adoption. I largely disagree with your assessment that it's user friendly, I've spent the last month fighting with my Linux router just to get built in features to work properly for one reason or another. But that's another story and I'm sure is related more to my incompetence with that distro of Linux than it is to anything else.

However, touching on that point (my incompetence with Linux), I'd say I'm largely representative of the non-Linux-professional crowd... even with my nerdclination. If something is that much of a pain in the ass for non-professionals to pick up and use, it just won't see such widespread adoption.

I'm really confused here, you talk about problems with Ubuntu and being a representative of the average Joe, but something tells me the situation is quite the opposite.

I entered the Linux world with no knowledge of it at all and had no relevant education. My computer expertise was building them myself and Windows general use. Ubuntu 10.04 at the time was my first experience, and while it was unfamiliar to me the (old) gnome desktop gave me an easy time of it, imo.

I wasn't trying to do anything heroic, I wanted a desktop that did the stuff I wanted, and it did. I should think that more represents the average Joe.

That said nowadays I hate Ubuntu, I feel it brings little of what it is supposed to bring on the software side. When you have a problem, the OS is so cluttered there are multiple solutions that may or may not work for any given issue, that sets the entry bar far, far higher than for instance getting into a desktop environment from the ground up in Arch.

As an advanced user your mistake was going Ubuntu, not going Linux. It isn't the easier distro, it is only the one where you have to do the least from BIOS check to a desktop environment. That is not a hard thing for others to catch up on.

Is there anything concrete in terms of 'drivers' and 'flaky', 'more fundamental features' which you can point to as being a problem for you?

Personally I've been using a bleeding edge distro (Arch Linux) for the past 5-6 years or so and I've only been bitten once by a stability issue with networking which forced me to downgrade the kernel while it was being fixed upstream (this would not have bitten me had I not been using a bleeding edge distro btw).

And I certainly don't have to spend all my time 'tied to the computer to get shit working', since Arch is bare-bones it certainly took more time to set it up exactly to my liking, but that's not something I have to spend time on anymore.

Instead I spend far too much time here on HN :D

I could write a laundry list of things, but like you, I'm too busy wasting time on Hacker News... when really I should go back to trying to research how to fix them :P

Actually, I've got a couple of things that flaked out on me in the last year or so. This was with Fedora 17, on a Lenovo Thinkpad x230 laptop, and Intel video. At first I put Ubuntu on it, but it wouldn't reliably survive suspend/resume, and sometimes X11 would lock up (this is because the chipset used in this laptop was band new at the time, and driver support was still a bit flaky). So I put Fedora 17 on it, and got a bit better luck.

However, after a few automatic updates, I started to get really weird sound issues. When playing videos with VLC, I would get a buzzing sound (like a broken speaker) for the first minute or so, then it would clear up -- this only happened sometimes, after a suspend / resume. Reverting to an earlier kernel fixed it, but also I couldn't duplicate the issue with any other audio program -- just VLC. Eventually (after a couple months), the various libraries, software components, and kernel versions stabilized on this hardware, and now suspend/resume, audio, etc are all flawless. But I know I'm probably in for some pain next time I upgrade hardware (unless I get last years model).

Second problem -- I picked up a generic 7.1 USB audio device. The center channel and subwoofer are reversed. I can't figure out how to re-assign the audio channels through the driver (I think there is some combination of Alsa and Pulse config file tweaks, but haven't got them to work yet). Ideally I'd like to pass some parameters to the kernel module for the sound device to swap channels. I looked at the source code, saw there was a spot where I can add the chipset in the "quirks" table, but couldn't find the data structures that assign audio output to specific channels. I'll probably end up jumping on one of the kernel dev mailing lists to figure that part out when I get time. Meanwhile, I just made an adapter to swap the channels (both the center speaker and sub-woofer come out of the same audio jack on the sound device).

Most of the things you have problems with are related to you not buying a machine with Linux on it. This is the reason SteamOS is so important.

We use hundreds of Linux machines, some macs(some of them are 5years old or more). We have no problems with it. Why, because we buy hardware that we know works with Linux. If it does not, we choose alternatives.

This solves all the problems in the future in our experience. We had no regressions as we could have with Windows(drivers that work on Windows X but not on the new version). This happened with the drivers of the few Windows machines we have with bar code scanners, we can't update the machines without changing the bar code system(big hassle), so we did not.

I couldn't disagree with you more about drivers that work on Windows breaking. MS actively tries to keep driver compatibility between releases. The Linux kernel developers actively try to break the binary drivers. This means that once the manufactures stop supporting the hardware, you have to worry about the drivers not working on the next kernel release. Case in point my Radeon X1950 can no longer run the catalyst drivers in Linux and I'm instead stuck with the buggy and slow Gallium3D drivers. In Windows 8, I can still load drivers written for Vista and run modern games like Star Craft 2 and Crysis 2.

If Valve can improve the driver situation, it will be a huge win for Linux gamers. Edit: typo

> MS actively tries to keep driver compatibility between releases

vista

> Case in point my Radeon X1950 can no longer run the catalyst drivers in Linux

wat. this has literally. nothing. to. do. with. linux. devs. AMD. does. not. support. these. cards.

in fact, your example of the X1950 is a perfect example of which cards the OSS community supports better than proprietary.

> in fact, your example of the X1950 is a perfect example of which cards the OSS community supports better than proprietary.

Wow, that's some twisted logic. AMD doesn't support the cards under Windows 8, but I can still load the unsupported Vista drivers and play modern games. Last time I benchmarked OpenArea in Windows vs the Linux Gallium3D, windows was 3x-5x faster, while the last working catalyst was 90% of windows. Beyond the Quake2/3 engine games, Gallium3D are worthless for gaming. I've had terrible luck with Humble Bundle games like Penumbra and demos like unigine that run perfectly under windows and Linux with the catalyst drivers, but are broken under Gallium3d.

The X1950 is in my spare PC. My primary rig is an AMD 5970 and I dread the day I'm forced to use Gallium3d drivers because of some GPL zealot kernel developer changed some driver interface. I'd love to see Valve pull a Google here to enforce a stable binary driver interface.

Stable internal ABI has been beaten to death. It isn't going to happen, nor should it.

It's very hard to get people behind a vision in a loosely decoupled system. People are going to implement their little things, solving their own problems, achieving a local maxima but it takes vision to create a coherent system. And look, Canonical is trying to do that and the reaction they get is bitching about MIR and other technologies they introduce. I wish somebody out there had a strong enough leadership to align all those efforts in a coherent product. I'm hoping to see a strong desktop leader emerge at some point.

Also have you considered that the promise of customization that you get from the various Linux distros is what makes you unsatisfied ? If you choose OSX or Windows you don't have much choice but to adapt. I remember when first using OSX how frustrated I was by the window manager but over time somehow I got used to it. With Linux I would have gotten into a hunt of the perfect (tiling) window manager, which then wouldn't work with system notifications and there goes another hour to find a solution. Then fix GDM or something else because my customizations don't work with the rest of the system... And every change introduces a new learning curve. Then you're not only a user but also a system designer.

I'm pretty sure you mean "loosely coupled," but I like the neologism.

Actually no. Linux has very widespread adoption from computer users and very little adoption from appliance users. That is because many of the things appliance users do with their computers (print things, play videos, listen to podcasts) require a specialized knowledge and an ability to parse user forums in order to work on Linux. There is no one to call up when Ubuntu doesn't print, there is no one listening at 1-800-PULS-AUDIO to answer questions about setting your mixer so your headphones turn off the speakers when they are plugged in. And perhaps, the most damning is when you buy a product that is used with a "computer" it rarely says on the package "Works with Linux" while it does say Windows and OS X.

Why not just accept that those people, the people who could care less about whether or not they can change their window manager or if Qt is better than GTK2 will never want to use Linux . They don't want to use a computer, they want to run "apps" which do something they want (play a game, or calculate their taxes, or help them write a letter). The more the underlying complexity of the system is reduced to zero the better, which is anathema to 99% of the FOSS community so it never ends up in the release.

> require a specialized knowledge and an ability to parse user forums in order to work on Linux

And don't forget, nearly unlimited free-time with which to do these things. Most working professionals don't have that, so whatever Windows/Mac device they choose is so they can spend what time they do have getting shit done, not screwing around for 3 or 4 days trying to figure out why the usb port doesn't work or how to maximize the value out of the $300 video card they bought without mucking around with recompiling drivers or some such.

Having recently started using Windows after years of using Linux exclusively, I'm confident that you can only make this comment with a straight face because you're intimately familiar with one system and not the other.

I feel just like you do, except IMO it's Windows that is baffling and unreliable.

Can you expand on this with examples?

Somehow after installing clean and using the system for less than a day I got into a cycle of BSoDs (maybe caused by a buggy RAID driver?). This required a reinstall.

Sometimes my monitors go blank and won't come back on. I still can't get a second external monitor working. I had trouble with the installers for various MS products, which produced completely useless error messages (e.g. unclickable URLs in dialog boxes that, when typed in, redirect to a generic marketing page). The standard process for finding and installing third-party utilities is a bad joke.

My point isn't about these specific problems, it's that the idea that Linux is uniquely flakey and unsatisfactory won't fly.

(I know the hardware is OK because I was using it with Linux and no problems a month ago.)

i concur. i'm a longtime linux/mac user, and I find some things on the windows machine I keep around for steam just really baffling.

As someone that works cross-platform I can help:

1, Copy and paste with the command box is a huge pain on windows. 2, You have to have focus on the window to scroll (my Kubuntu means I can hover over a window and use the scroll wheel, its so intuitive when work on one document and reading another). 3, Windows Explorers URL area is selected from the keyboard shortcut Alt-D as opposed to IE/Firefox/Nautilus/Dolphin/Chrome which is Ctrl-L. 4, Updates on windows can take half a day. 5, Configurations are split between a magical windows registry and the users profile folder. Outlooks nk2 file is an awesome example. 6, The black box problem. I have found a bug in a windows component, however I have no way of getting a fix or knowing if someone is looking at it. All I got last time was that my report will be forwarded to the appropriate team. 7, Painfully slow start up 8, Reboots after updates. 9, Error messages which don't explain much and problem helpers which provide nothing extra. 10, Windows Help Documentation!!! 11, MSDN which is little more than reformatted header files. How many complete (working) examples of code are there in the MSDN. Now compare that to the Qt documentation.

Man. we could play cliche bingo with this post. You confuse a distribution with the kernel, say devices don't work (cos it's still 2005), or exclaim that it's just too hard to learn. Apparently, you've never written a program in C in your 20-year programming career, either. Perhaps someone could do a kickstarter to fund a better trolling template.

but I'll assume you're genuine in your opinions. Try kubuntu 12.10, then come back and tell us how much you hate linux again. I dare you.

Sorry, I wasn't trying to play cliche bingo as you so eloquently put it, nor was I attempting to start a religious war - because I don't honestly feel there is a religious war worth having. I think there's plenty of room in the market for Linux, and I think it deserves to be there. Desktop just needs its awkward corners knocked off and polished up a bit.

...and yes, I have programmed in C in my 20 year programming career - albeit closer to the beginning end of it than the present end of it... that doesn't make Linux any more user friendly.

...also, I never said I hate Linux, I don't hate it at all. I am just finding it remarkably frustrating as I don't find that anything is easily discoverable without a ton of digging and reading reams of information and blog posts - if, indeed, you can find the relevant items. If I hated it, I wouldn't be on my 4th distro in 2 months, I'd have given up and gone back to Windows... or bought myself a Mac.

I'm glad he approves of it. Generally if he doesn't approve of something like this then it triggers a string of "Torvalds says" articles that do nothing for nix. Then someone brings up a RMS angle and then it all devolves into an artificially created shtstorm which keeps the entire FOSS scene in this constant paralyzed state.

Personally I think Valve - as great as they are - will step on a pile of RMS flavored toes. So be it. They bring raw power to the (desktop) nix world like never before. nix hasn't really moved an inch on the desktop front anyway...whats the worse that could happen? The risk is acceptable given the massive potential progress.

I think it'll also provide much needed focus to the nix world. Sure variety has its benefit...but at some point it becomes fragmentation not variety. A central vocal point like valve will go a long way to fixing that.

nix hasn't really moved an inch on the desktop front anyway

What a ridiculous thing to say. Millions of people use *nix operating systems on their desktop computers every day. This number is significantly higher than a few years ago. When I hear someone complaining about state of GNU/Linux on desktops, they usually mean they can't play gamez. Sometimes they add something about lacking good video drivers (to play gamez).

And by your statement you reduce the whole range of possible computer uses to gaming. Maybe we shouldn't confuse "linux in entertainment" with "linux on desktops"? Former is only a subset of the latter.

Don't get me wrong, I like playing, and I adore Valve and their games. But for many GNU/Linux users giving up their freedoms may not be worth another brainless AAA title.

I'll complain about it, I'm not a gamer... Ubuntu won't even suspend/sleep properly under LUKS for crypting out loud!

It has moved on the desktop front - a few inches anyway.

I use desktop linux as my primary OS every day. It's still not ready for prime time, but it's finally close enough that I can deal with it.

Some fun things that happen to me on a regular basis - you'll note that these issues have nothing to do with "gamez":

    * my sound stops working and/or diverts to anther output.  I still can't get it to detect when I put in a mic or headphones with any reliability. Yes, the hardware is compatible. 
    * behavior on attempted suspend or wake is always a crapshoot.
    * it shuffles my desktop layout every time I unplug/replug in my two external monitors 
    * memory gradually gets consumed. Even doing a restart of X doesn't fix it. (No, I'm not talking about cache.)   Under mac I was rebooting once a month. Under windows it was once a week.  On *desktop* linux I do it every 2-3 days just to keep things working.  (Server-side linux goes for months.) 
    * sometimes networking stops working. I have to disconnect/reconnect wifi. 
    * invariably, on a daily basis, I find I have to tweak *something* to make it work or keep it working.
    * Here's a fun one: using vim with clientserver enabled resets my Konsole appearance profile (fonts, colors, transparency) every time I open the editor or remote-open a file with an existing session.

Obviously none of these are deal breakers - the positives outweigh the negatives. The control, user interface, tools and projects available, and many other things make the tradeoff worth it to me. But I wouldn't call it a smooth experience, or even an experience on-par with other desktop OSes if we're talking about ability to "just work" without issues.

"But for many GNU/Linux users giving up their freedoms may not be worth another brainless AAA title."

Neither you, I nor any other users who are intensely interested in freedoms are the target audience for SteamOS. SteamOS won't affect what you and those users and I can use Linux for.

And - because I suspect somebody will pipe up with some veiled-snark solutions or suggestions for my issues - you're missing the point. When using a stable desktop OS, I should not have to spend time on these things. I didn't have to on either Mac or Windows.

>What a ridiculous thing to say. [...] This number is significantly higher than a few years ago.

Significantly higher yes...off a low base. We're talking 1..2..3...maybe a (naive)5 percent of progress on total market share over literally decades in the desktop market. Compare that to the androids absolutely crushing the entire mobile market in a few short years. So yes in the bigger scheme of things "nix hasn't really moved an inch on the desktop front".

>And by your statement you reduce the whole range of possible computer uses to gaming.

You're kidding, right? I focused on gaming, because we are discussing a post about Valve. The effects will extend to every corner of GUI driven nix given that nvidia has already made real world changes based purely on this announcement. The nix world can fight another 10 years of a losing war against an unmotivated nvidia & linux or they can grab something like this by the balls.

>gamez

Really?

Compare that to the androids absolutely crushing the entire mobile market

Why would I compare totally different things? When Android came by, there was no monopoly in the mobile OS market. Every vendor had its own. Also, there are much less legacy applications than on desktops. And mobile OS is substantially easier to catch up with. Prior to smartphone era users had no trouble switching between phone vendors (and subsequently the OSes). Switching to a new a desktop OS is way more difficult step.

Better compare it to OS X. Same decades in the market, and still only ~15% share. Yes, more than GNU, but they do their own hardware, and OS is by far not their main selling point.

I focused on gaming, because we are discussing a post about Valve

Sorry, I shouldn't have jumped on you so quickly. Yes, GNU/Linux seriously lacks in a few more areas besides gaming (namely, professional media editing). But it is still a far stretch to say Linux isn't advancing on the desktops. Jeez, 10 years ago we didn't have a decent office suite.

Anyway, what makes you so concerned about Nvidia motivation? My integrated Intel video does an excellent job drawing whatever 2D I throw at it. And it even can do 3D well enough to play some 5yo games (if you consider TF2 one of those).

The "potential progress" was already there. It's just been held up by proprietary companies every step of the way. In the case of games, nearly every single game company refuses to provide their games as free (libre) game software. In the case of GPU drivers, the two big players still refuse to provide free drivers. There are plenty in the free software community willing to contribute to graphics drivers and to porting games to GNU/Linux, but these companies won't let them because it conflicts with their goals of exerting strict control over their own users. I haven't seen any evidence that this situation will let up any time soon because it seems that people have fallen for their PR tricks and so don't seem to mind whether or not most drivers and game software are proprietary. And so the cycle of abuse continues.

The whole reason why we use free software like GNU/Linux to begin with is because it doesn't make us subservient to one vendor like nonfree software does. It disturbs me that I see people making statements implying that "we need this particular proprietary company to come along and save us" or "we only need to be patient, slowly but surely they will port their games to GNU/Linux" because this is exactly the kind of thinking that free software discourages. As long as these companies are in the business of publishing nonfree software, you can be guaranteed they will drag their feet every step of the way. People need to stop buying into their lame excuses about why they can't publish free software ("We can't make money that way" is probably the oldest and stupidest one), excusing their abusive behavior and then begging them to do things only makes the situation worse.

>It's just been held up by proprietary companies every step of the way.

With that kind of angle the FOSS movement might as well fall on its own sword.

Profit driven companies have zero incentive to bother with FOSS & thus they don't bother. One can whine about how unfair that is (see Torvalds vs Nvidia). Write epic arguments about it even...but ultimately that achieves exactly nothing.

Steambox changes the game fundamentally though. There is now a real economic incentive here for nvidia etc to commit real resources...and that has the side effect of benefiting pure FOSS too.

Yes RMS, FOSS, half a percent gain in nix usage, the year of the linux desktop and all that crap...the real world does not care. nix has to take some real world risks to get anywhere...and no moving from gnome 2 to gnome 3 is not a risk...despite all the drama it caused.

>In the case of games, nearly every single game company refuses to provide their games as free (libre) game software.

Obviously. Sinking a couple of million into a game and providing it for free is not exactly a winning strategy. Or worse, devs do so and then introduce microtransactions and similar evils.

The current model suits me perfectly (aside from the DLC bullsht). Once off payment for indefinite access (well until obsolescence anyway).

They don't bother because all their users have been convinced not to care, perhaps by proprietary software marketing efforts. The free software movement seeks to educate people on why they should care, and why it's valuable for companies to give users freedom instead of trying to lure them into proprietary licensing traps. I don't know what you gain from viewing GNU/Linux as a product that is supposed to compete with proprietary software on the same level, because it's not that at all. As we've seen from countless failed companies trying to pitch GNU/Linux on the desktop, traditional marketing strategies for proprietary software do not carry over to free software. This is the true misallocation of resources and it's the only thing causing the "wheel-spinning" you're referring to.

>Sinking a couple of million into a game and providing it for free is not exactly a winning strategy.

It's free as in freedom, not free as in price. Most if not all games sell based on content anyway, which doesn't need to be free (as in freedom or as in price) to resolve the problem of the users being treated like garbage and being forced into a subservient relationship.

>In the case of games, nearly every single game company refuses to provide their games as free (libre) game software.

Rockstar spent $266 million to develop Grand Theft Auto V. How do you expect them to recoup this cost if they gave it away as free software? What communist fairly land do you come from?

If GP is from the RMS school of thought, then only the code needs to be free. The assets aren't. The code, with no assets, is practically worthless to the end-consumer.

until the Darkmod Standalone was released...

>nix hasn't really moved an inch on the desktop front anyway

My copy of Mavericks begs to differ.

The only reason I chose Windows over Mac OS X for my new home computer is I wanted to play games on it.

If SteamOS lets me play AAA games (and not just Portal, Half-Life, Left4Dead and Team Fortress) AND it lets me run XBMC or Plex or some other TV software, then I'd switch in a second.

Pretty sure it will, if not we can def make it do that. I am so looking forward to this.

I'm certainly not a "gamer" any longer, but I still have a pretty healthy Steam library.

I've been very happy with the increasing options on Mac OS X and I hope they keep improving.

I'm way more excited about slower, turn-based games, and Civ V, Europa Universalis, and FTL have been enough to keep me from searching for anything else.

Despite this, I know I've been missing out on games I would probably love. I think I see a SteamOS device in my future.

If you've already bought the machine and Windows OS then you should be able to install steam on it, buy their new controller and have a roughly similar experience with the big difference that the whole steam library will work today, without requiring streaming from another box.

SteamOS will be great for linux, but people seem to assume Valve are taking a very exclusive, traditional console approach, when they're trying something a bit different.

I also assume that basically any Steam game that comes native to Linuxt will also come to Mac OS X.

"The desktop" that is relevant for GNU/Linux distributions is a stagnant, shrinking market. You will see more Chromebooks than you will ever see Linux desktops. At first glance SteamOS does not even look like a GNU/Linux desktop. It's supposed to be a living room machine, complete with its own "SteamBox." Attributing anything important to SteamOS is akin to saying "Roku is Linux; this is really going to help us get some drivers."

I'm guessing you don't count chromebooks/ boxes as "Linux desktops"?

No, I'm not an RMS disciple but here I said GNU/Linux for a reason. Chromebooks aren't GNU/Linux. Torvalds doesn't seem to be saying that Steam will help with Chromebook-like devices--and why would it? The manufacturer will do what it takes to get the drivers working on those devices.

If you use just "Linux", then there's some ambiguity. If you use both "GNU/Linux" and "Linux" to refer to the same thing in consecutive sentences, then that's plain misleading.

>Chromebooks aren't GNU/Linux.

It does use glibc and many other common software found in GNU/Linux systems. I'd say it's a GNU/Linux system.

If they help sort out some decent graphics driver code I will love them forever and send them cake on their birthday.

No cake necessary, just buy their games. :-D

Yeah, the cake was a lie.

shh...

You monster.

I think it will help not exclusively _Linux_ but it will help more to get real support behind graphics drivers. Sure, some of this is directly related to the kernel (like power management, switchable graphics etc.) but mostly it is stuff which isn't directly related to the kernel (e.g. mesa, xorg, wayland). What I see recently that AMD has really stepped up too (if you own Radeon HD, not the very old Xwhatever series), contributing to kernel code and mesa drivers. Nvidia has some pretty good proprietary drivers, but heh... they are proprietary. Thats the main problem I have: Both AMD/Nvidia closed drivers offer better 3D performance, better feature support for their chips etc. but when I have trouble with graphics on my system, it's because of the closed drivers. You never know if they still work after updating.

I know this isn't the point of the article, but:

> developers focus on useless UX features

This is the number one reason I have personally delivered poor quality software.

It makes me wonder how different is this one from the dozen of Linux distributions around. Will this one solve any of the followings problems? if so, it's a very good start.

Linux is hard to use on day-day. Even you chose the most friendly distribution (like Ubuntu) Almost nothing works in the first try to install. Look, I failed to get the newly gcc version (I really want it) because I need to compile from source code and a lot of libs is missing from my system and when I try to install others lot of libs is missing too. It's get in looping. I got the latest version of Microsoft Visual Studio in just one second. Where can I get something better? Also, nowdays everybody is switching to a XBOX/PS4 etc and no longer do much play on PC. I don't have numbers, but I can seen it's changing a lot. It shouldn't decrease, just unlike, it means in a future (very close) playing big games on PC is almost dead. Why should someone invest time and money in something going to no longer exists? PC is something that the peoples will continues to have to work and browser on internet, if not on your smartphone/tablet.

GCC 4.8 is in the repos. Interesting that you had trouble getting it.

Show me how to install it?

Reply on news.ycombinator.com