SteamOS, Linux, and Steam Machines
steamcommunity.com
steamcommunity.com
Those who were there at the beginning of the Valve Linux Offensive (circa 2012) will remember Gabe Newell stating that the main driver for investing in Linux is their dependency on Windows as a platform. Valve has never been one for dependencies at the best of times (creating their own engine to replace quake, creating their own sales platform to replace retail) and having a dependency on a platform that explicitly seems to be trying to emulate the iOS/Android approach more and more is absolutely a reason to at the very least have a plan B.
Having Valve come out and say that they're still invested in Linux is important, but what is highlighted in the post is also poignant - in 2012 it seemed the barrier to adoption was interest in the platform and access to a Linux machine. The solution to both of these was to create a high-profile, easily accessible form factor which was Linux by default. It turns out that this may not have actually been the right solution. Linux adoption has risen and the back catalogue of available games has increased incredibly, despite lacklustre sales. The problem now, and perhaps always was the graphics APIs and cost of porting. Valve are actively investing in Vulkan which is absolutely the right thing for them to be doing. The only downside is that it looks like they're neglecting the 'movement' they started with Steam Machines.
I hope to see big things from Valve in the future. If I were them, however, I would invest in creating a truly fantastic desktop experience with seamless transition from Windows, lots of toys to play with, riced desktops etc. Hardware and the console form factor should be secondary to creating a PC desktop experience worthy of the people who created Half Life and steam.
Still, it is a bit sad. Steam machines could have worked, but they were severely flawed. Buying a steam machine did not guarantee that one would get a machine actually capable of playing current games, they were very different and often underpowered. Linux-compatibility alone is then not enough.
I think they could be great if they had been more controlled: Define a form factor, define the minimum cpu and gpu power needed, buy those in big quantities to be able to give a discount compared to buying a PC manually. Release it at the start of a gpu-release cycle, so it does not need to be updated that soon. Pre-install SteamOS on a machine like this and add a starter game (Portal) and one would actually have an upgradeable gaming console developers could target games for.
Maybe throw in a pro version of the hardware for people with money to burn
So a common hardware target would do absolutely no good. No games would be developed specifically for that target.
Just because the market is a certain way doesn't mean that's out of necessity. There are large markets in developing economies that would benefit much (as would everyone else) from getting out of MS's grip.
Steam Machines are the failure. Making them even more expensive would not have helped.
The lowest end Alpha was by all means a decent console contender in terms of size, graphics, processor, and memory, while also being one of the cheapest models of all Steam Machines.
The majority of OEMs pushed out overpriced garbage, but for the most part Alienware wasn't one of them. They committed to a custom, low profile case design AND worked with Nvidia to create a powerful low power GPU based on the GTX 960m. $400 for that versus the other companies pushing poorly supported low end Radeons? Phenomenal.
I bought on release day and still use my Alienware Steam Machine. Unfortunately, like many other said above, the developers simply don't optimize for Linux releases like they do for Windows, but the hardware pushes what you'd generally expect from what is effectively a couple years old gaming laptop hardware.
Is Valve as a company actually ever capable of throwing it's weight behind anything?
They had VR handed to them on a plate and all they've managed to ship is a tech demo. Almost feels like all these projects that should be major for them are treated like side projects.
I know people still have a lot of love for the games they released 13 years ago but for me I've learned to accept that Valve today isn't what Valve back then was.
They own the distribution of VR games. Even games that can't be distributed on Steam use their APIs, because it's the widely-supported way of doing it that works with all the headsets. Even if Valve never makes a VR game, they're going to make a lot of money from them if it ever takes off.
They're just an online store, and for people like myself who grew up on Half Life it's a sad realisation.
I see a problem that is just as big, or probably bigger, here for the (PC) game industry: the near-monopoly of Steam.
I try to buy non-Steam versions of the games I buy, because I dislike Steam's online DRM and their client, but it seems that Steam is the new Windows when it comes to gaming. "Everything" is released on Steam first, and it seems a lot of times only on Steam. Which means that we're one focus shift or stolen password away from not being able to run the games we've bought and paid for.
gog.com has a better offer, IMHO, by being DRM free. Are there any other similar competing services? Humble Bundle seem to be selling Steam keys these days, so they're a subset of Steam.
UPlay, Origin and whatever the hell Blizzard's client is called these days are just lock-ins from their respective publishers.
There might be others sellers, but I don't know if they deals with the distribution themselves or just sells Steam keys, like HumbleBundle.
(No affiliation)
They mostly have games by Japanese developers; I suspect their Japanese version of the site has quite a few more games, however they do have some games with English translations that do not appear on gog. I haven't actually purchased from them yet, mostly due to their cloudflare settings being fairly strict and often blocking my vpn/proxy (that only I use, but is in an IP address range that others use for vpn) so I am reluctant to use them even when it works. However, it is not so good if gog is the only DRM-free option either (I strictly avoid steam). I don't know of any others.
Humble occasionally has bundles that have "DRM-free" options for most or all of the games (like the current indie bundle), although I am a little suspicious that their "DRM-free" might sometimes have call home DRM. However, I haven't investigated closely and most of the games I got from them have worked fine without an internet connection.
Itch.io
Many games you get from Steam can be run without the client.
itch.io is one such competitor, with lots of interesting games (you need bit of digging for sure). http://itch.io/
And it wasn't exclusive, they sold other people's games on there too. Then they helped kickstart the indie gaming explosion with the Potato Sack. Then they started trying to make Linux a better gaming platform to provide an alternative to being locked in to MS. Then they released a non-exclusive VR platform against a competitor that was exclusive.
Say what you want about Valve having a monopoly and that being bad for the market, because that's generally true, but my opinion is that Valve has demonstrated a sincere interest in doing good things for PC gaming, not just what's good for Valve.
Valve has invested a lot of time and effort into its various loot box schemes (hats, trading cards, stamps, badges, etc), and Steam is currently the only retail platform deeply complicit in the worst slot machine tendencies of the industry at large, and has been for a long time.
It's been 6 years since their last franchise title, CS:GO, was released and it's profitability was based on micro-transactions. I think this experience (along with DOTA 2) would forever change the way the company perceives its value to customers.
In hindsight, I regret hats on TF2. I regret that as customers we steered Valve into this lackluster state where microtransacted eSports is in and beautiful game engines for storytelling are forgotten.
What a waste of time.
None of the big name releases ever seem to be available. I understand this is due to the game studio's decision of not spending the money on porting to a market that they calculate would make up a rounding error in revenue.
I'm sure there are Linux-compatible indie games out there that would be nice to try, but I'm a casual gamer and I don't have time to research and trial and error what might be a good fit for me.
A large amount of AA and Indie games are also released on Linux (largely, I suspect, because they use middle-ware like Unity which makes multi-platform releases much more straight-forward).
My point being that there are lots of options if you have a hankering to play a quality game on your Linux machine. The problem in this instance, I suspect, is more to do with this line from imjustsaying's post: 'if I have that rare desire to load up a game'.
Specifically for first-/third-person games this means getting the right sensitivity and disabling mouse-acceleration (or setting it in a very specific way for some hardcore quake pros supposedly).
Last time I checked, disabling mouse accelleration is a huge hassle and no working UI exists. The average gamer will not go into config files of the OS to try and find the right formula that makes their mouse move the way they expect it to.
This isn't really an important issue in normal desktop usage or point-and-click games. Thus support by linux developers is understandably lacking.
man xsetEven somebody who is familiar with the concept of man pages would have no chance of knowing that xset is the one to look for.
Discoverability is super important. Luckily, the kind of desktops people actually use tend to have a System Settings app in which mouse acceleration can be figured.
Unfortunately it looks like the path from "apropos mouse" might be a bit more convoluted.
xset m 0/1 4I've found it a little harder to script in ~/.xsession though for hotplug devices. Things get addressed by integer ID, not necessarily names.
Perhaps some special financial incentives from Valve to big studios would be able to get the rock moving down the mountain? Temporary discounts perhaps?
They seem to be going low-moderate risk/investment. Sure there's progress, but then the Wine devs are moving pretty steadily too at the moment.
If Valve wants to really be serious:
- Offer incentives to Devs to support Linux/Vulkan
- Release a goddamn new game (HL3 - dreams) and make it Linux exclusive for a time
- develop and release new features for Linux first (maybe if GoG ever get the Galaxy client over to Linux Valve will get a kick in the pants to act...)
- Promote SteamOS and (maybe) Steam Machines, like at least some. Show Steam Machines as an 'out-of-the-box Windows gaming replacement, not a console competitor.
- Fund Linux devs or take a lower cut from Linux releases
- Convince Adobe to release their suite on Linux! Push for/fund official Unity and Unreal Engine Linux versions
- Offer the ability to natively launch games through Wine (and either use 'community' environment files or require one exist, or have a link to WineHQ for help)
Be serious Valve, be genuine and get your flat structured 'follow the shiny thing' company focused and determined.
I think people seriously overestimate how popular that game would be. If they really did release it as a Linux exclusive, a lot of us nerds(and I mean it as a compliment here) would install Linux in a heartbeat just to play it, but I think overall, the game would sell very poorly. It has been so many years since the last one that the mainstream public doesn't care about it anymore - a group of hardcore enthusiasts would buy it + some people interested in what the whole Half-Life hype is about, but I'd be surprised if it sold million copies.
Let's assume millions did rush to buy it and install Linux.
Think of how many of them would be installing it just to play one game. Most would gripe about not being able to do something they're used to in their other OS [0].
While I love Linux, there are some pain points on desktop. Just getting it installed on your machine is not always trivial even when you know what you're doing. Something you've done a thousand times sometimes just doesn't work for no real reason and you just have to try again [1].
Linus help them if they need a proprietary driver [2].
I imagine most of the people trying Linux for the first time would be trying it with the intention of playing a game instead of learning a new system. That sets them up to be annoyed every step of the way and walk away angry at the new system (Linux).
On top of all that, it could cause some frustrations for people like myself who try to help others install and learn Linux. I imagine having several conversations with friends/family thinking they want to finally learn some of this Linux thing I keep talking about. Only to discover it's just because they want to play HL3.
[0]: http://linux.oneandoneis2.org/LNW.htm
[1]: I realize this happens with computers, in general. The point here isn't to bash Linux (heh).
[2]: obviously, not a fault of the Linux ecosystem, but that doesn't change the fact that it is incredibly frustrating.
EDIT: formatting.
If memory serves, when Valve was initially getting their feet wet, publicly, with Linux, they were getting Left 4 Dead (1 or 2, can't remember) to work there. If I remember correctly, they got it to a point that hardware for hardware, they got the Linux port to run better than the Windows version.
And that's a direction worth going if you want the gaming crowd to follow: come this way and get an extra boost in performance. We know you just spent $300 to get a couple frames per second boost, come get a few more for free.
They really need to do something about Big Picture mode though: that thing looks terrible on anything above 1080p. Even more so at 3440x1440 where it is blurry and inside black bars. If they're trying to make Steam into an appliance of sorts, the user interface needs to be right.
He tried CS:GO in my Linux box and liked that it just worked, apparently ran smoother, and according to him the OS itself seemed simpler. He wants to try Linux on his computer now with my help. Let's see how that will work out, but still, it's a start.
I'd like to eventually be using an open source operating system on my desktop, I really would, but Linux Desktop is crap and from what I can tell will remain crap because their community has crap ideas.
But now Windows is getting more crap too, by trying to copy the crap parts of the mobile ecosystem and the crap parts of Linux at the same time.
I'm facing a future where I end up with a Linux Desktop not because the Linux Desktop has finally become worth using, but because Windows has become so shit it is better by comparison.
It's enough to make me want to give up computing altogether.
I'm curious what MS Windows has copied from a common Linux DE or distro that you believe is "crap"?
Yeah, that's a big part of the problem right there.
> I'm curious what MS Windows has copied from a common Linux DE or distro that you believe is "crap"?
Rolling releases, i.e: break everything every 3 months, an over-reliance on the the command line for configuration tasks (and I say this as someone who loves powershell), package managers, and hardcoded paths (mostly for their dev tools), to name a few.
Things break on Ubuntu for those addicted to upgrades (me!).
In my preferred DE, KDE/Plasma, I can't recall ever needing to use command line for normal user level configurations.
Which in typical Linux evangelist fahion means it never happens to anybody else.
But sure, there are edge cases where you might need to use the terminal or edit a configuration file.
As an analogy, in Windows I've occasionally had to make configuration changes through regedit.
That doesn't mean it's normal to have to make changes in the registry on Windows, it can usually be done through the gui
It depends on your hardware as much as anything IME - sadly most PC hardware sellers make stuff to work under MS Windows, so if you're not careful with your purchases you can have problems.
I used to use Slackware, and that was a lot of command line - or at least text based - configuration; but I chose it to learn about administration. Like buying an old car so you can practice fixing it.
Kernel support is improving too. AMD is contributing to the open source amdgpu driver now, Intel's open source driver is one the most supported among distros.
And what if a Linux Desktop ends up being better than Windows at some point in the future, well, why won't you use the best tool you have available?
Windows is often a horrible thing for people to learn at first, but they do it because it's the default.
If Linux is to be in the sphere of 'things to learn to play games/be productive,' then a more disruptive event is better than hoping to get there by growing at 0.2% per year...
Of course, he also still asks me every year when "the next Half-Life is coming out".
- Convince Adobe to release their suite on Linux! Push for/fund official Unity and Unreal Engine Linux versions Maybe they should just convince Microsoft to port over win32 and all the other APIs to linux? Whilst they're at it, they could also convince your most hated political figure to be more like your most loved political figure. With the convincing powers like these, Valve could just convince the world to no longer suffer from hunger.
The only incentives that work would be of the financial kind, however, paying for support is not sustainable. Console vendors can pay for first party or exclusive support from game devs because they are fairly sure this will increase revenue in the long term. Valve gains no more money if they incentivize developers to support Linux. I also don't really want for things to be supported exclusively on Linux unless if there are good technical reasons - I use Linux because of the freedom it provides me, I wouldn't advocate for people to be forced to use it. FWIW, Unity and Unreal run on linux, but a lot of developers elect to use windows-only Unity plugins. The developers don't care.
What Linux really needs is non-shit drivers and a mainstream desktop environment that takes into account input latency when developing it's window manager - Wayland on Gnome is really bad in this regard, yet Gnome is the target desktop environment for Steam OS (at least last time I checked). They should maybe make sure that their games run just as fast on Linux as they do on other platforms - currently this is not the case. I believe that solving these issues will be hard and may require a unilateral effort across various entities in the ecosystem - something which is incredibly hard to do, especially when it comes to Gnome. Then there's things like drivers - yes, performance wise Intel and Nvidia drivers might be OK, but they are not feature complete especially when it comes to power management. I don't believe there is a single problem that has to be fixed to make gaming on linux a massive success - there isn't even a shortlist of 10 incredulous tasks. It has to be a long-term effort and the open source community has to take part in it.
That statement is not fair to the people who do this every single day. We're only where we are because the open source community took part of it and took us there, give some credit where it's due.
edit: don't mean to sound ungrateful, point is OSS is dramatically underfunded, we need big corps to do more
Valve has only something like 300 employees last time I checked, so you can't expect them to do a lot more unless they massively hire a lot more people.
You meant it as a joke but by far and large hunger is a problem of the past. Massive famines have not occurred in years, and where they do it's because of political disorders like conflicts or civil wars.
https://footprint.globalance-bank.com/_Resources/Persistent/...
There's also increasing problems inside first world countries: https://policy-practice.oxfam.org.uk/blog/2014/06/uk-hunger-...
It's also possible that hunger is a leading indicator of conflicts. I've heard this cited as a cause of the Arab Spring, and water conflict is a factor in the war in Yemen.
The capacity for ending hunger is here.
Probably UK companies throw out enough food to feed all those who're hungry in their locality [1]. But hunger is still a big problem. It's going to get worse here too. The recession brought food prices up significantly, Brexit will do so to.
[1] high-street companies used to sell off perishable goods cheaply at the end of the day. Now they lock them in bins. Because poor people would buy them to save money, poorer people would "steal" then from the bins.
B) There's currently at least one massive famine going on RIGHT NOW in Yemen that gets pretty much no attention. Yes, it' due to a conflict, but that doesn't mean it isn't happening.
I used to toot this horn, but I'd rather see the Affinity line up released, at this point. Aside from professional designers, and people who use it daily, the costs are beyond prohibitive.
I worked as a designer for about 15 years before landing a job in programming. So, my uses for Photoshop or Illustrator have dwindled down to once a month at most. A subscription makes no sense for someone in my shoes.
I understand why they did it, and how piracy was out of control for their apps. But $50 a month for something I'm rarely going to use is nuts. Which, is a real shame considering the amount of hours I spend learning these apps.
Wait, what? Valve has been making huge contributions to the large parts of the Linux ecosystem.
Of course they can do more -- anyone can, but patronizing them with "be genuine" and "follow the shiny thing" is really unnecessary.
It's an exhausting and unfair reality.
In short: "The Copenhagen Interpretation of Ethics says that when you observe or interact with a problem in any way, you can be blamed for it. At the very least, you are to blame for not doing more."
[0] - https://blog.jaibot.com/the-copenhagen-interpretation-of-eth...
Just because someone (or some group) contributes something to the Linux ecosystem doesn't necessarily mean it's a net positive. It could, arguably, be a negative: it could cause more fragmentation and confusion, it could make Linux look bad to outsiders (decreasing adoption rates), it could take mindshare away from a better organized project, etc. Worst case, it could be done with malicious intentions, such as contributing code that has a secret (and not obvious) backdoor/vulnerability.
Also, I think there's a logical fallacy about criticizing people who don't contribute yet offer criticism, though I can't name it. For an extreme case, imagine a group of lousy coders writing an OS kernel, and some actual kernel expert like Linus or Dave Cutler jumps in an offers some criticism, though they aren't contributing to this particular project. Who should you listen to, the actual expert in the field, or one of the amateurs who really doesn't know what he's doing? Obviously, the expert's criticism should be considered and not rejected out-of-hand just because they're not a contributor.
Dunno, my Steam Box runs games like Grid Autosport, Mad Max, Tomb Raider, Deus Ex etc. without a single issue, as fast as on Windows (quad i5 + NVidia), and I don't even notice I am on Linux. So thanks to Valve and Feral Interactive, I can slowly move away from Windows to Linux for all my computing needs, not just for programming, Deep Learning and blockchain. What else is missing?
Moreover, I can run a lot of software under Wine, like Adobe CS6 or MS Office, or even many games (it has gotten significantly better lately).
It runs them now, but none of those games ran on Linux when they first launched. Grid Autosport took 17 months after the initial release, Mad Max took 13 months, Tomb Raider took 3 years (and the 2015 sequel is still not available on Linux) and Deus Ex MD was the only somewhat timely one at 2 months (but the previous game in the series was never ported to Linux at all).
I know not everyone cares about playing games the moment they come out, but many people do. Especially with multiplayer games where the community is most active closer to launch.
You might consider saving that money instead of promoting Linux through buying games you won't play.
Compound the interest, and maybe you can find an indie game studio investment instead. That would be a real and substantial contribution that moves the needle.
https://www.gamingonlinux.com/articles/looks-like-the-linux-...
And then, there are runtime considerations like where you store the player data, cache files, and the case-sensitiveness of the filesystems which many Windows developers are completely unaware of.
So no, Vulcan doesn't give you "zero effort" Linux ports.
however...
If you release games on a platform, or targeting a hardware spec which has a player base with is too small (like say, VR), and don't make enough money in your 'big bang' release, you're screwed, and your company goes under.
Heck, even if you do make some money, if its not enough to 'wow' investors, your studio will probably still be shut down. The economics of big bang release software are very hard ('Explain to me again why after making a million dollars this month you want another million dollars investment that will not generate any return for the next 2 years?'), and game companies are behind the curve in dealing with it.
The alternative (freemium) is just a race to the bottom, and its happening on steam right now the same way it happened on the app stores.
Incentives to encourage developers to target 'linux first' isn't just not going to work; its actually unethical.
You're basically paying people to target a platform that you know isn't significant enough for them to recoup the significant development costs in making games.
So yeah.
Realistically, you need to do one thing:
Figure out a way to provide a meaningful continuing income stream to developers targeting the platform.
As I understand it, what value is currently doing is attempting to bootstrap the market to consume linux games, by:
- Minimise the cost of developing on the platform by improving the tooling and making it easier to target 'cross platform' to include an additional platform (vulken, steam networking, etc).
- Make a platform that will consistently actually run said games, and push that upstream for other distributions to pick up.
- Work with partners to provide hardware that is configured in a way that can actually run that platform without having to compile the kernel yourself.
So... frankly, that seems like a pretty smart way to tackle the problem.
...but on the other hand, it hasn't been massively successful, to be fair.
Practically speaking, 'linux first' is never going to work; but I don't really know what else they can really do to help it along the way.
Care to explain?
People seems to think this kind of politics from first parties and even from Microsoft very acceptable. But for some reason these tactics applied to Linux would be unethical?
I think the prospect of developing a game that "automatically" works on both Windows (and not just 10, but also 7 and 8) and macOS would be quite compelling. Plus the bonus that the games would also work on Linux, and iOS/Android, and eventually the web (at least some of them for these last three).
Besides Apple, Valve should also convince Sony to support Vulkan on the next-gen PS5 console. They should tell both of these companies that they benefit more from being in this larger ecosystem than they do by creating their little walled gardens while Microsoft's gaming ecosystem engulfs and dominates them from outside.
It's also quite imperative for Valve to try to achieve this because they are still at risk of being obsolete by a future version of Windows that restricts Win32 apps by default for consumers (a non-zero chance of that happening over the next 10 years). So Valve should have a big incentive to try to get some kind of "partnership" with both Apple and Sony and get them to support Vulkan.
Metal is low-level enough that Vulkan can be implemented on top of it without too much overhead.
Most game companies don't care about Linux, because the added development costs on testing and community support don't translate in any meaningful revenue.
They aren’t trying to promote Linux. They are trying to have an open platform to run and sell games.
Why would they be talking to Adobe, limit their sales while pissing off fans by releasing a Linux exclusive (for any amount of time,) or spend/forgo money on Linux development/sales?
This doesn’t make any sense. It’s just a wishlist for someone who likes Linux, not a part of strategy to deal with DirectX.
While saying Linux First outright is probably a bad move, they do need to funnel developers' attention elsewhere to something either they control (a linux distro called SteamOS) or something compliant with Linux in general so that whatever flavor of the month distro can run them. I'd personally aim for the latter as it enables sales to people in love with Arch, Ubuntu, etc., but I could see a push for SteamOS too.
But yeah, Linux only? Not a good idea. Creating incentives for companies to target Linux too seems like a good path, but at the same time, you have to make it worth their time to support another version.
This is about DirectX, and not about what you use as a desktop OS.
A move like this would definitely not win Linux any friends, because as it turns out the Linux Desktop really isn't that great and you'd just be causing a huge headache for a lot of people. The gamers would be annoyed at having to deal with Linux's bullshit, the Linux Destkop community would be annoyed at the huge influx of new users who don't think "try another distro" (especially after the third distro), or "write your own driver, it's open source!", are acceptable solutions to their problems.
If Linux wants gamers there's a simple (note: simple != easy) thing they have to do: make their platform attractive to gamers and game developers. So far they can't manage either, the only reason gaming on Linux is even remotely viable is A) Valve coming in and saying "Ok, since there's no such thing as a standardized base system in Linux Desktop land, here's a standardized base system for Steam you can target", and 2) the popularity of Android making engines like Unity and Unreal have Linux support, which allows devs to tack it on without too much extra effort.
That's not true. Stick to the 'user friendly' popular ones (Ubuntu, Mint, etc) and they're very solid now. Sure, edge cases are still a problem, but for gaming - if supported - not really anything more problematic than Windows.
When did you last try Linux?
I guess it's a matter of habit and expectations then?
The camera fails to work consistently (the driver does not always load)
Audio fails to work consistently (restarting pulse fixes this)
Closing the laptop lid initially sleeps, but then the laptop wakes up a few minutes later for no reason (so I wrote a script to disable all the events that can wake the laptop up, but this in turn causes it to not wake up at all sometimes when I actually want it to)
kde plasmashell will crash if I try to use ffplay, and doesn't reliably recover so I end up needing to restart my session which loses any work I hadn't saved since I can't switch windows
These are all problems I have never experienced with my Windows laptop, nor macOS.
Anecdotes obviously, but I've been trying to go only Linux for the past decade and never felt like I could reliably do so.
It's important to buy hardware that is supported and properly documented by the manufacturer. Apple is not one of those that do that.
Perform vanilla install of Windows 10 on new desktop build
Watch it sporadically work with Logitech 920 webcam
Watch it randomly not load network device at boot
Watch it show me a blank login screen with no ability to pick a user or enter pwd
Watch it “lose” a USB connected Brother laser printer
Watch it ...
I’m pretty sure Windows and Apple get a pass on quality because we’re low level conditioned to give Big Corp products a pass
The Ubuntu + i3 setup I’ve been rolling for years now has not given me any more or less odd ball problems than Win or OSX
OSX on my work laptop reboots every day. I have Firefox and Slack and some stuff installed via homebrew, on a 2017 MacBook from work. Every morning “your computer is recovering from a shutdown”. Even if I shut the software. IT hasn’t figured it out, Apple can’t and won’t replace hardware that tests ok
Watch FaceTime just stop doing anything every call I make with it
Oooh and the WiFi issues with Mac that I haven’t had with Linux since wpa_supplicant was managed by hand
But Apple has their “just works” rep. I mean it “just loads” then just needs a reload everyday seems more accurate
I’m not saying Linux isn’t a quirky chore
I’m saying it doesn’t seem any quirkier than the competition going off my daily experiences with all 3
Starting from my Commodore128 till today, other than that original Commodore, I haven't seen anything else as well-built as Windows. macOS & desktop Linux distros included. I hate to elevate one above the other but my honest opinion starting on desktops in 1986 that Windows is the most rock solid desktop OS out there today. Microsoft has to ensure it's the most battle tested by necessity, given that even back in 2011 it ran 1+ billion[0] devices. There's nothing else even in the same league for desktop operating systems when it comes to exposure, testing & fixes in the QA feedback loop. Even the big player, Ubuntu, doesn't come close.
[0]http://www.businessinsider.com/right-now-there-are-125-billi...
With ThinkPad I never had an issue.
It's nothing like it was 5-10yrs ago.
The person you're replying to is right, using Linux is a lot like using MacOS, you need the right hardware for the software. I too had issues with Macbooks and it was like night/day switching to Dell/Lenovo when using Linux. When everything works, desktop Linux is very nice and on the same tier (or better) to MacOS as a programmer.
If you want a stable experience, you'll need a lts release with supported hardware. Eg for Ubuntu: 16.04 along with a Dell xps. Or red hat, debian stable or whatever.
Fortunately Ubuntu 18.04 is just around the corner.
Not saying poor hw vendor support for Linux isn't still a problem. But my parents have actually had more issues upgrading to Windows 10 (firewire drivers not working, won't be fixed by vendor or ms).
Seems to be trend that when hw doesn't work on Windows, it's "pc doesn't work/is difficult", on Mac it's "well, third party, can't expect it to work". But with Linux it's always: "Linux is crap".
I said the Linux desktop experience has not been good to me, with any of the hardware (even the stuff I picked specifically for good support) with very popular hardware. I still use Linux as my desktop OS, and really would love to ditch Windows 10 permanently but don't feel like I can without a lot of frustration. The same as it has been for the past 20 years for me.
- Works flawless. - No issues with the webcam. - Plasmashell don't crash. - Even I can use my old Epson C-62 printer where on Windows I can't!
And I have nearly the same configuration on my workplace without issues.
The only issues that I notice: - Skype for linux is garbage. - OBS doing weird thing the last time that I try, I need to update it.
Works perfectly? Exaggeration isn't going to change anything. "Those anecdotes are bullshit but mine aren't, they were doing it wrong" is a common trope in these discussions.
At least with laptops, there are models that you can research and see if other people have been successful using a Linux distro on them.
Building your own Linux gaming rig is then that much harder in comparison for most people.
I really want Linux to be the go-to OS that people install, but the criticisms of it tend to be pretty valid, IME.
Sleep function works better that Windows. I don't get blue screens when I wake up, like windows does. Also, the time since I boot to get ready to do something, is far better on Linux that on Windows.
Also, I can output sound at same time by HDMI and to the sound card out put. I can't do that on Windows!
And Video card support, with AMD, it's at the same level that Windows.
> And Video card support, with AMD, it's at the same level that Windows.
I haven't used Windows in years, but if that is the case, Windows must have gotten way worse. I can't say my experience with any video-card vendor has been smooth on Linux.
Someone posts legitimate complaints, someone else comes out of the woodwork saying "works for me!", dismisses those complaints, and wonders why nobody gives a shit about their desktop platform.
My recipe for success lately is: use all Intel hardware for major components (CPU/GPU/Net), use Business grade hardware (I run Thinkpads and Optiplex desktops at the moment), and stick to a rolling release distro or otherwise frequently updated one (I use Arch, there are probably plenty of other good options, but don't try to run a 2 year old LTS distro on 2 month old hardware if you don't want a headache).
Wifi: randomly drops out, packet loss, inability to connect.
Audio: I still end up having to disable Pulsaudio about half the time.
I generally use older Thinkpads and Debian, though I've used Arch, Ubuntu, Mint and probably others, as well as a range of hardware.
Intel and Atheros wifi works great out of the box 99% of the time
Video Card support: video drivers are still a pain point for sure, but I don't think it's fair to say 'video card support is terrible' as a blanket statement. Intel, Nvidia and AMD graphics all work on linux.
AMD is making great progress with their OSS drivers, and intel has been well supported with OSS drivers for a long time now.
Most issues are due to nvidia still insisting on only supporting proprietary drivers, and the annoying issues that come with that (e.g. drivers breaking after kernel updates).
Audio: Usually just works out of the box
Sleep: In my experience sleep works fine on most hardware, but there are some motherboards with horrifically buggy acpi implementations that have issues.
Or Ralink. I bought a TP-Link Archer with Ralink that had a driver and was a recommended AC chipset on some sites at the time. The driver loaded and worked, but frequently dropped connections or had very slow transfer rates.
AMD graphics, NVIDIA
Still can't reliable switch my monitor on/off without problems (artifacting or miscellaneous Wayland compositor bugs) with the amdgpu driver. I used NVIDIA with Nouveau before, but that was even less reliable (regular flickering). At least amdgpu works fine if I keep my screen on and avoid system sleeps.
Nevertheless, I can't help but laugh at the fact that you basically just suggested for him to try another distro.
More seriously, the reason that suggestion is so common is because it does sometimes work out and I can't blame you for pointing that out.
They released HL2 as a Steam exclusive (IIRC), and your description of Linux now was how you'd describe Steam then (not great, a headache). It won Steam lots of friends and set it on its money-printing way.
I'll be the first to agree that there isn't much on the line for Valve now, given that Microsoft has pulled back from the more extreme plans of Windows-Store-exclusivity fuckery, which was an existential threat to Valve.
Having once been a Linux Desktop evangelist, and knowing what I do about modern Linux Desktops, I feel I can safely say that very few people are going to be won over by it.
There is a ...tiny... difference between "install an additional shitty drm software" and "replace your entire system". And yes, even steams introduction was hard, back in the day with shitty internet... i certainly wouldn't spend that much time (and money for isdn traffic) installing a game these days.
After installing it, the machine would hang during boot. I had to boot to recovery mode to remove the driver. Then I tried to switch to Xorg from Wayland. That also caused the UI to fail to load. I eventually figured out how to get Xorg to work and tried the AMDGPU driver again. Again, failed to boot. I ended up reinstalling Ubuntu from scratch. I've consigned myself to the fact that I can't resolve the tearing issue and I'll just have to live with it.
So, I dual-boot Windows 10 and Ubuntu, depending on the task I'm working on.
Why is everybody who is complaining about the Ubuntu issues trying to run what amounts to a beta build of the operating system? I think you would have a lot smoother experience of the just a to 16.04 and don't upgrade until everything is ironed out.
I have a Linux machine at home and I use it regularly. Linux is great as a Unix system. It has a more modern userland than macOS. Linux is exceptionally good for development (from recent gcc/clang to profiling with perf). But why can't we just be honest: Linux and hardware is hit and miss. You need to buy the right hardware and configuration typically requires at least some terminal work and/or configuration file editing, which are both outside the realm of regular users and many gamers.
When they've already committed to purchasing and assembling special hardware components just to run the games? Crosschecking your build against Linux compatibility would be pretty simple...
After going a bit back and forth on canonical/Ubuntu - I'm starting to think I've slightly misunderstood what they're doing: they really are all about competing with red hat on a debian foundation. This is both good and bad; Lxd and juju seem great, and they have a live kernel patching service (costs money :-/), support contracts... And put effort into various projects that benefits everyone.
But they're not really trying to be a great free software project/product - a "better" debian than debian.
I think that's actually for the better - but they've got the signalling a bit messed up. Almost no-one should run non-lts Ubuntu as their daily driver, IMNHO - just like Debian testing is for the adventurous - not for the person new to debian.
But every Ubuntu release comes with friendly packaging that understandably confuses newcomers. And by the time people have few years experience, they've become tinkerers - the target audience. So they stay, and pull people on to the edge, when most would be better off with the boring stuff that works pretty great.
Maybe it's just poor/understaffed community management. Debian has a wiki that's half dead in places (I've made some half-hearted attempts to help give it live support - but it's though keep going when there's so few around...) - While Ubuntu have vibrant user Web forums ripe with enthusiastic and often very misguided advice...
Every time I mention the macbook, I am always chided that "it's a lemon." In that spirit, given the preponderance of people who use linux daily and have no issues[0] with it, I think it's time to stop claiming the issue is linux itself already. It isn't a valid argument anymore in 2018 and given the fact about drivers on linux, it's even more surprising how well it does when one has driver support.
[0] No issues that wouldn't be something out of the norm of what windows users or mac users are already used to regularly dealing with.
I suppose it could be a hedge against Microsoft doing something incredibly stupid, like cutting support for Win32 and mandating UWP garbage through their garbage store.
I completely disagree with that. I don't believe anyone should be forced to use a different platform, when the only reason behind it is a handshake and a piece of paper - even if it's my preferred platform.
Kinda goes against one of the reasons for using Linux.
It's been a long while since I launched a game on Steam; I was under the impression that a Windows binary is required.
"So we’re going to continue working with the Linux distribution guys, shipping Steam, shipping our games, and making it as easy as possible for anybody who’s engaged with us — putting their games on Steam and getting those running on Linux, as well. It’s a hedging strategy.
I think that Windows 8 is kind of a catastrophe for everybody in the PC space."
> https://venturebeat.com/2012/07/25/valves-gabe-newell-talks/
Given Gabe's fears never really materialized (perhaps with the exception of Windows 10 'S', but even that no longer has fees to enable third party software sources), there simply hasn't been anything like as much pressure for Valve to push SteamOS/Linux.
So the incentive here is to continue to make it viable in order to have an argument against someone trying to blackmail you because you are so locked-in into their ecosystem.
On the other side, the more viable you make it, the better future you are building, as you then have access to a platform where you can actually control from end to end, and so implement whatever your future requirements require you to.
They are not in a hurry to change the status quo, they are playing it for the long term.
I run UE4 on Linux without errors.
But everything else you mentioned I agree with 100%. I'll also add my opinion that Valve is lazy and hard of hearing.
Vulkan, RadeonSi and VR to name a few. They deserve all the recognition they can get for it.
It seems that it's even trivial to release Stream games for linux when using any common game engine, and I hope developers will make use of that as much as possible.
Fallout 3 has no linux version in steam unfortunately, but I actually could play it perfectly by running the windows client of steam in wine. Everything just worked (I had to do a memory patch to support 64-bit CPUs, but so do windows users have to so no difference there)
Playtesting takes time and money and it's entirely possible that just exporting the game for linux won't run perfectly, so I can see why some companies using the common game engines don't do it, I kinda wish they did, though.
Wine runs a lot of Windows games properly.
The two major commercial producers of proprietary operating systems are basically giving away their operating system at this point. Free Linux distributions offer basically the same functionality for most users; the only things preventing migration from one OS to another are existing applications. For non-gamers, that's often moot, because the world exists in a web browser now. Yes, there are other applications causing OS lock-in too, but they're mostly on the business side, not the consumer side.
There is still plenty of money to be made from enterprise operating systems, but the innovation there is in a totally different direction than what consumers would find valuable.
Example: Linux containers are great! But who's spending the R&D money to make them a viable way to distribute games? There's plenty of work being done on containers and graphics chipsets for ML applications, but not for consumer use.
AppImage, snapcraft, etc
Huh? Windows 10 home and pro aren't free? It's true that Apple doesn't charge for upgrades, but they do charge a significant markup - and they will happily sell you a licence to run Os X on a hackintosh although they'll forbid you from doing it in their licence ("non apple hardware").
> Until a business model emerges which challenges this
Appstores. Windows, and Os X have app stores. Steam is an appstore. Microsoft and Apple bundle a competitor to steam with the Os.
Just because both apple and Ms suck at making/marketing/developing their App stores... Doesn't mean valve don't want to stay in the business...
They can't get in on ios, Xbox, ps4, etc. They could probably make an appstore for Android - but I'm guessing they'll just focus on streaming games from the cloud to Android (esp Android TV).
With hw support for virtualization, Linux could be the universal runtime - allowing the same game to run on Mac, Windows and Linux...
It's probably because they don't pay their developers and fire everyone that expresses any discontent with it... KSP is not a good citizen.
I've been playing older titles that now have official Linux versions, like Bioshock, Metro 2033. Let me tell you, I would never have predicted I'd be playing AAA titles natively on Linux.
Valve’s hardware team has been very busy, designing the Vive in 2016 and the knuckles controller in 2017.
They’ve said they are working on multiple full length VR games, but that they wanted to be able to design the hardware (knuckles) and the games together a la Nintendo.
This is public knowledge.
It’s also known that Oculus is releasing two standalone VR devices in the next couple years.
Valve has a headset, controllers, and a Linux OS. It seems obvious to me where this is all going: a stand-alone Valve VR device running SteamOS.
Of course they’re not investing in screen-based games and desktop Linux boxes. They’re perfectly positioned to release a medium-defining device in a green field new computing paradigm.
They are at least as well positioned in this fight as Microsoft, Facebook, and Apple. That’s the big leagues. Valve makes good money on Steam, but not that kind of money.
And you know Apple and Microsoft aren’t going to use Linux. So having Valve means there’s a Linux horse in the VR race.
I personally have been using Linux on work computers and personal laptops (almost) exclusively for.. I don't know, 10 years? - but my "gaming pc" is still on Windows and I'm not in the mood for experiments with WINE or grasping at the games that run on Linux.
Does anyone play major games on Windows and Linux? I'm really excited for the idea of Linux gaming, but it's not a thing I'm willing to compromise and fiddling if the games I want to play really work.
>While it's true Steam Machines aren't exactly flying off the shelves
I hope they didn't expect anything else. Steam already has a limited choice of games and a console that is even more limited is a niche product at best.
To make it succeed, they need a lot of resources to get GPU manufacturers and game publishers on board (also those who took their games off Steam and created their own launchers). They need both of that only to have a chance to attack regular consoles.
I have been a proponent of gaming on Linux most of my adult life, but I can't help but think that Valve might be missing the boat by sinking resources into gaming on Linux local hardware. Times are a-changin'.
Valve's goal was to hedge against a "Windows Store" model for games at a time when it seemed that it was inevitable.
Also, and I don't mean to sound dismissive, but that streaming model simply will not work. When it comes to anything more than visual novels or turn based RPGs, the only way you'll be able to solve the problem of latency, which players of these games are very sensitive to, is by using tachyons. I think there are better uses of hypothetical faster-than-light particles, personally.
> I have been a proponent of gaming on Linux most of my adult life, but I can't help but think that Valve might be missing the boat by sinking resources into gaming on Linux local hardware. Times are a-changin'.
I thought this when OnLive launched, but even with a relatively good internet connection for the time, OnLive was still pretty horrible to play consistently. You can have a great experience for an hour or even an hour or two, but by the 8th hour you're bound to be hit by the inherent problems of delivery of a network. Maybe that's a random latency spike or loss of video quality. Maybe that's something more. Even a great connection can have problems with consistency over long durations. That's what you need for streaming games to work well for everyone.
There is absolutely no reason for me to have a Windows machine except the games I play that don't work on Linux/Wine.
But it has the first mover advantage and that is really everything. Even platforms that are infinitely better, have better tools, support, and pricing (such as itch.io that doesn't require a cent to publish your games) are heavily disadvantaged just because it takes extra work for developers and gamers to adopt them. There's no easy way out of this and Valve can continue being lazy and make major slip-ups every year and there's very little others can do to endanger their heavily entrenched position on the gaming market.
As far as PC (or GNU/Linux in particular) gaming is considered, Steam is everything, and it's not going to change any time soon.
We'd have an open handheld platform for indie games.
Let's be realistic, the success of the Switch is more because of Nintendo software (Zelda, but also the promise of other Nintendo IPs coming) rather than the hardware. The success of indies on the Switch is more a byproduct of the Switch success that the other way around.
Nvidia Shield existing before Nintendo Switch, it did have a focus on gaming but it didn't cought anyone's attention. I don't see a partnership with Steam changing much about that.
They already had a standalone Shield years before the Switch. Nobody wanted to support it since games had to be reworked for ARM and the Android framework. Same with the Ouya. It's not a particularly difficult obstacle, but there were so few users that it wasn't worth it.
Breath of the Wild alone was enough to sell millions of units. There are now over 15 million Switches out in consumers's hands.
Nintendo is the last company in the world that would be interested in this. They're not in the hardware business to sell other peoples software, they're in it to build a platform for their own software.
Nintendo platforms could literally ship with zero third party support and people would still buy them. Heck most casual gamers I know who bought Wii and Switch don't even own third party games at all, just Nintendo.
My point was that Steam should have been interested in it.