Is SteamOS officially licenced? Does that mean I can expect stuff to work just as well as on the Deck? With the Deck, I don't care what OS games were built for, everything just works.
Depends on the words used for the device. Valve updated their brand guidelines recently, this shows that there are certain devices that could be classified as fully verified by Valve to work: https://steamdeckhq.com/news/new-steam-brand-guidelines-hint...
Have you encountered extreme lag on your Deck? Most 2D games work great, but when I try to play Elden Ring or Stray or Jedi Fallen Order, it works well about 90% of the time, and the other 10% there is extreme lag that makes the game unplayable.
I've hit lag quite a few times with Hogwart's Legacy. Dropping the graphics quality down helps but even at the point it looks like shit it often still stutters, especially during complex combat situations (the point at which stuttering is least tolerable).
I love the Steam Deck and they do a really, truly great job at tuning profiles of most games to run well, but the most demanding AAA games do sometimes hit a perf ceiling.
If you're curious about the details and you haven't seen it before, you can enable metrics/stats in settins and see a screen overlay of your performance. As someone who frequently keeps htop open on my laptop while working, I appreciate this feature :-)
Played all of these without any major issues. I've only seen lags for the games I installed on the memory card, so I put small indie stuff on there and keep AAA on the main hard drive.
The first game I played in Linux was Arkham Asylum, which would be unexpectedly slow in certain levels (like the botanical garden). I haven't experienced that in other games.
I'm curious to see how 3rd parties with SteamOS handhelds behave.
Both in the context of the handhelds themselves (do they meet the steam deck's high bar for reliability and longevity) as well as how the companies handle things like open source contributions or vetting games.
It takes more than just the OS to replicate the steam deck's success (even though SteamOS itself is very well executed)
I have a ROG Ally running Bazzite. It's not Steam OS per se but close enough. The performance overlay works, TDP controls work (albeit it takes a little initial setup to install the plugins), fast suspend/resume works. Performance is really good. It was shocking to me how well Windows gaming on Linux works in general. Granted, I don't play online games that require anti-cheat software so that does skew things a bit.
Bazzite is the first time I feel like PC gaming on a TV is as convenient as a console. And you can put it on a 150$ mini PC from AliExpress and play pretty much anything that isn't an AAA game from the last 3-5 years or a competitive online game with unsupported anticheat, including most emulators. It's truly an amazing achievement.
Yep. Been running it on a 7000 series Ryzen with an M780 and it can run pretty much everything, including emulators, at good enough speeds and framerates - not just locally, but streaming to an Android handheld as well.
Games get vetted for Proton, which is a DirectX implementation for Linux. More powerful hardware might perform better, but whether or not a game can run correctly is a property of Proton, not of the Steam Deck specifically.
UTM for iOS has been basically neutered on performance so I doubt it would even run. The authors even had to rebrand it as a "retro OS emulator" because it can't run anything newer than XP.
I've never owned an iphone (and have no intention of ever doing so) so I can't speak from direct experience, but I would heavily bet that a far easier path would be to use the Steam Link iOS app to connect remotely to your gaming rig (and to be clear, a Steam Deck works just fine as a gaming rig if you don't have a bigger PC you'd prefer to use). If you're on the same LAN, performance is good enough that you may not even notice it's not running locally.
Wireless VR and Stadia are both insanely impressive. Think about how much data you've gotta be moving (and how quickly) to have an 8k video stream feel smooth, or have a game computed in a data center feel local.
It's starting to occur to me that even without having any time or inclination for video games these days, one of these handheld Linux gaming computers seems to be what I actually want out of an Android phone, and I'm really starting to despise the whole phone ecosystem. Does one of these exist with passable front/rear cameras for video calling/spur of the moment shots?
I never even use my phone as a phone. I certainly don't make calls while I'm out and about. The thing pointlessly enables spying on me when I don't even want it. But a pocket computer with wifi, enough storage for my entire music collection + wikipedia + maps, and ability to take phone-quality pictures? Running an actual Linux distribution so I don't run into stupid issues like how Android doesn't let you use NFS so they can push you into their cloud services? Awesome.
This one doesn't and neither does the Steam Deck. Generally these are packed to the gills with screen (no good space on the front for a camera) and cooling (taking the space that a back camera could use.
This is the first third party steam deck, but you could just use any USB camera. None of these are really pocketable. You can hook up a raspberry pi or any similar SBC to a screen and peripherals and make a DIY Linux phone.
A desire for a portable but fully capable computer is mostly why I bought one of the original Steam Decks. You can trivially plug it into a dock and have a full and pretty capable "desktop" experience out of it as well.
At that point I didn't even own any Steam games and hadn't played a game in well over a decade, and had no interest in doing so. I'm a sucker for open hardware though. Forward a few years and I've been really happy with it. With games so easily accessible I have bought quite a few over the years now.
Adding cameras would be awesome! Though were I in charge of product for it, I'd want to think very carefully before doing so. It is a general purpose portable computer, but it's important to remember that the vast majority of the market doesn't buy it for that, and a camera and associated software may get in the way of it's primary purpose... Though, it could enable some pretty slick live video chat during multiplayer, which could be pretty awesome for people that play those kinds of games.
Yeah sometimes I wish Valve would appy some of that micro transaction effort into just building a phone + steamdeck controller wrapper. And then start capturing micro transactions on mobile.
Part of what's exciting about SteamOS is that Linux on a touchscreen is a disappointing experience otherwise. There are a handful of UI options (GNOME, KDE Plasma Mobile, Phosh, Maui), and they're all works-in-progress to varying degrees.
I've been using Jovian-NixOS as a placeholder for SteamOS, and have been disappointed at how mouse-and-keyboard-centric Linux UIs still are.
Not entirely what you are asking for, but Waydroid has a surprisingly simple architecture which makes it easy to tinker with to suit your needs. It just links the Android hardware compositor OEMs use straight into waylands interface, which is made easy since the interfaces are already very similar to each other. It's basically no different than how it runs on your phone once it's up and running. Non-wayland inputs/sensors work in a similar fashion, basically just getting the output of the /dev file then redirecting it into the android equivalent.
All it's really doing is pulling a LineageOS image, then dropping in the waydroid "drivers/firmware" that make it work with wayland using a series of patch files people have contributed. It all then gets run in a container which you "remotely" access for the GUI parts. Starting and stopping the container is like turning your phone on and off. If you set it up as a systemd service then you can just make this lockstep with your actual boot sequence.
It's primarily "marketed" as a tool to run android apps on Linux, which means the documentation kind of guides you down this path of running individual apps as their own window in Gnome or something. This doesn't work very well and apps tend to crash when you try to resize them. However, if you run it via `waydroid show-full-ui` it works perfectly in my experience and operates just like a phone/tablet.
I currently use it with a 2-in-1 x1 Thinkpad (meaning it has a touchscreen + pen), primarily for reading comics, books and watching movies when I connect it to my TV. I have it "wrapped" in Hyprland, so I can easily switch workspaces to return to the "host" if I want to, but I usually just do everything entirely within waydroid, and lazily use JuiceSSH for most host operations.
My laptops camera is a bit special and lacks linux support, but my understanding is a properly supported camera would work fine with waydroid. It doesn't support screen rotation or wifi configuration though, so those parts still need to be handled by the host system. I have my laptop auto rotate via hyprland since it was the easiest but I think it would be trivial to just add a patch for support as well.
It uses a pretty old LineageOS image, but I rebuilt it with some community patches so I could use the latest image, so you aren't even restricted to just LineageOS. I also didn't quite like how my stylus was being treated by android, so made a custom idc file for it and plopped it into the overlay filesystem waydroid has, saving me the effort of having to rebuild the image again.
It all sounds kind of complicated, but you can get most of this up and running right out of the box, complete with google play services too (requires you get a token from google, but they have guides for this). It's just once you notice something bugging you, you can actually do something about it, in an afternoon even, instead of just dealing with it forever. Something I personally appreciate.
That said, I'm using it as a tablet, for a phone there will be some gaps that need still need filling, such as:
- You would need to somehow get a small touchscreen device and install linux on it
- GPS: You would need to make sure this device exposes the GPS interface in some way and integrate it with waydroid. Since waydroid works via the same path OEMs take, there are probably some guides for this.
- LTE: Same as GPS, though so long as you can get internet on the host, waydroid will inherit that automatically.
- ARM Apps: Waydroid has some stuff that should allow you to run ARM apps but I've never bothered to set it up, not sure how well it works.
What should work out of the box, assuming they support linux:
- Cameras
- All touch interactions
- All stylus interactions
- Keyboard + Mouse
- HW rendering, works perfectly
- Network access (only configurable on host though)
- Direct USB access (out-of-the box in the sense you need to enabl...
It has been clear for several years now that non-Android Linux is the only way forward, but short of big changes like the EU stopping acting like a coward and banning both iOS and Android (or, yeah, maybe Valve trying their hand at it), progress is very slow (though still faster than a few years ago).
Why, just why won't any of the competitors make one of these handhelds with at least the same amount and types of input options as the Steam Deck. I'd love a higher performance handheld but none of them have two trackpads, gyro input and four back buttons, so I'm stuck with the Deck.
It's an extra cost that doesn't add much to the device for north of 90% of their intended audience. I love my Steam Deck but barely use the trackpads except for the rare times I need the keyboard or I wind up in the desktop doing something.
The Go S appears to only have one rear button on each side.
I've played half a dozen games on the original Go, and I don't think I've used more than one rear button. In Half Life, it's more comfortable to hold for crouch. In one of the Spider Man games, it expected a trackpad click, which was easier to perform with a rear button.
Since games have been designed to be playable with only the face buttons, there's not always obvious utility for extras. Maybe as premium controllers become more popular, this will change. For now, I can definitely see the argument that 4 buttons are superfluous (except for standardization).
Just because everything is bound to the face buttons doesn't make it the best solution. Developers make do with what they have. Extra buttons mean I don't have to move my finger off the trackpad or joystick every time I want to do a common action, particularly while aiming the mouse cursor.
I have a Deck and almost never use the track pads, they feel totally redundant to me. The right one is pretty good on desktop mode and in games that aren't Deck-optimized and need a mouse, though. What are you using both of them for?
I use them to play Stellaris/Rimworld/Oxygen Not Included. Rimworld/ONI you could do with Joysticks but I'm just quicker with trackpads and it's finer tuned control.
I pretty much use them constantly. They're useful for playing any shooters or anything with mouse look, they're great for RTS, city builders, point and click adventure, Factorio, etc. Currently I'm playing Dawn of War, Ground Control, 688i Hunter Killer and Fleet Command, which basically requires trackpads. I also play Stronghold Crusader, Warhammer Gladius, Company of Heroes, Warframe, WoW and other MMOs on it. I honestly don't know how people get by without them, they're extremely useful.
Even for emulators, they're great for binding all kinds of extra functions.
Using the right trackpad also means I can free up the right joystick for other stuff, and there's always enough other stuff I want to access regularly enough.
XInput is still the standard API for controllers on PC, and there's no way for it to support additional inputs. So this is the same issue with adding gryo and extra buttons to any controller: they don't work without software to translate the new inputs to those standard either XInput or keyboard and mouse events.
Like how ds4windows is used to translate Playstation controllers gryo and touchpad. Even the Steam controller requires Steam. I don't think oems are investing in the software to do the same.
This, the Lenovo Legion Go S—Powered by SteamOS, shares the hardware as the Windows variant, Lenovo Legion Go S. They're not building hardware specific for SteamOS yet.
I think the Orange Pi Neo is the only (other) one that's targeting Linux only. And they do have touchpads
It's TBD how it will handle extra buttons. I believe Valve has a kernel module for the Steam Deck controller.
Lenovo's current handheld exposes an Xbox controller and a USB device with extra buttons. Handheld Daemon synthesizes them into a virtual PlayStaton controller. If you tried to use the current SteamOS without Handheld Daemon, the 6 extra buttons wouldn't work. The community is waiting to see if Valve ships a more robust solution that Linux can standardize on, or if we'll still be using daemons like HHD for devices that don't have official Valve support.
Yeah, but that requires Steam to add support for those input events. It doesn't have an API for you to feed it input events. There's not exactly a standard for random input events for Steam to hook into either.
That's why distros like Bazzite and ChimeraOS have an additional translator[0] to take input events and output a controller that Steam supports, Dualsense edge or Steam Deck's controller, to be able to add support for gyro and back buttons.
This is something to watch : we'll see how much Valve actually cares about Linux (and libre software/hardware in general) compared to just using it to get rid of Windows : will they make their new controller API SteamOS exclusive ?
(How well do Steam Link and Steam Controller work without Steam ?)
I'd say on PC because it's easier to name the modern controller protocols that aren't Xinput: Nintendo, Sony, and Steam. Except for Steam, the first two aren't targeting PCs. Basically every other controller targeting PC from the old Logitech F710 to the Flydigi Apex 4 will be using XInput to communicate with the computer. Even the Hori Steam Controller has an alternative XInput mode despite being the only current standalone Steam Controller.
Steam Link the streaming application does require the Steam. Steam Link the discontinued hardware was able to run local applications without Steam.
The Steam Controller defaults to a keyboard and mouse mode until it receives a signal from Steam. So even on Linux, the Steam controller required Steam to function as a controller until the protocol was reverse engineered. Can read about it in the driver comment header, and the Valve copyright that didn't get added until they contributed support for the Deck controller in these drivers : https://github.com/torvalds/linux/blob/master/drivers/hid/hi...
(I recently started again regularly using both the hardware Steam Link and the original Steam Controller - funnily enough, for a non-Steam game - I'll have to check out what people have hacked over the years...)
I've thought about this too, and given that Valve is brilliant when it comes to pricing, I suspect that they have data showing that gamers are well aware of the "Windows Tax" and expect to not have to pay it for a Linux-based device. I want to be conservative advancing that theory as confirmation bias is a real risk (since that describes me perfectly) but I do see little bits of reinforcing evidence around.
The way Microsoft killed EEE PC movement was by removing the Windows Tax, the moment OEMs no longer needed to pay for Windows on EEE PC style devices, they were mostly gone in one year, and the new tablets segment killed the remaining survivors.
Lets be real, with SteamOS having a Windows ecosytem, it is about saving Windows licenses mostly, and not really fostering a GNU/Linux gaming ecosystem.
My main gaming PC is one of those cheapy "mini gaming pcs" you get on Amazon for like $350. I installed NixOS on there, and have it automatically boot into GameScope with the SteamOS interface.
After I got everything set up (which wasn't hard), the amount of maintenance I've had to do on there is zero. SteamOS managed to make desktop Linux accessible to basically anyone; it was easy to build my own game console that I enjoy considerably more than my Xbox One that I never bother with.
Valve has made desktop Linux a lot more viable than I ever thought would happen.
If you mostly deal with AMD stuff, most things "just work" with Linux. In my case, WiFi worked out of the box, graphics driver loaded fine, sound driver worked fine.
I assume the chief reasoning behind this decision twofold. First, it's a crapshoot to get alternative stores working on my steam deck (I wouldn't expect my less technical gamer friends to feel confident installing a game from Epic/EA/Ubisoft). It's also a financial decision, as Valve presumably isn't charging the OEMs a licensing fee.
In a market that's getting more crowded by the day, where so many of their competitors are using the same SoC and come in at roughly the same price point, it's a rational decision to control costs as much as possible. Not to mention, they also need to compete against the Deck itself, which Valve is subsidizing to hit its pricepoint. The alternative would be to innovate and differentiate your product offering from your competitors (like the original Legion Go, which had detachable controllers and a fancy pseudo-mouse). The fact that this machine didn't carry those features forward is a strong signal that the bottom of the market is where the early wins will be.
Keep in mind also that Microsoft is planning their own handheld xbox experience. Speculation: MS' effort will look a lot like Valve's, with Xbox games coming day-and-date to OEM-built handhelds, as well as a first party reference implementation that's intended as a benchmark for devs. I imagine that most OEMs are using the cheaper steamOS pricing as a hedge against future gouging from MS; if there's an "Xbox certification process" that proves to be too cumbersome, a vendor that's already shipping cheap(er) devices en masse with SteamOS has the option of simply dropping windows support altogether.
It's a crapshoot to make Ubisoft games to work on Windows desktop PC already.
On one hand you have the shitty and/or DRM-plagued stores, on another you have DRM-free ones like GoG, Humble Bundle, Itch.io... which in my experience Just Work (tm) in a few clicks thanks to the likes of Lutris launcher/Proton/Wine.
(And in the middle, Steam(OS) to rule them all.)
P.S.: I've heard that Battle.net games also work easily and well, but I don't know how long this will last with Microsoft buying Activision-Blizzard and Valve ramping up their fight against Windows...
Like EEE PCs, this will only work while Microsoft decides to tolerate Proton and the way Valve runs Windows applications on a Linux distribution.
The day they decide it has been long enough, like they did with XP licence that ended up wiping EEE PC, Valve would have wished to have fostered a native GNU/Linux ecosystem instead.
Presumably the EU/etc. would go after them if they tried something like this? Besides desktop isn’t really as important to MS these days as it used to be.
> fostered a native GNU/Linux
Wouldn’t that require a stable ABI and such? Making Linux binary/proprietary software friendly seems like a much harder task than [potentially] fighting MS over the right to reimplement their APIs.
Why build something from scratch when Wine/Proton is already there and is pretty much the same thing?
Of course that would be the trivial (if still very complicated and expensive on its own) part, incentivizing developers to invest the amount of resources needed to port their games to a new platform (which would have single digit % market share at best) would be a lot harder.
Sure there might be some risks. But I don’t really see why MS would be bothered to actually try and kill Wine. The risk of antitrust etc. would be very high and MS wouldn’t gain too much from this anyway.
Besides.. the Supreme Court already ruled in favor of Google over the Java/Oracle thing not sure how would this be much different.
And? I fail to see how is that so valuable on its own.
I mean don’t get me wrong I do see some merits but realistically if Valve went that way they would struggle getting of the ground and it would just be a waste of time/money.
Possibly one pathway that could actually be feasible would be to leverage Wine to gain significant marketshare and then incrementally transition to a more “native” approach.
> the hard way.
How? I mean what could MS actually do about this? Besides trying to compete by offering alternative options.
Just the existence of steam deck is the transition to a native approach, they put pressure on devs and game engines to support, but proton is a weird middle layer, it would be easier to support Linux and vulkan natively and that's growing
There is zero pressure on the devs, as it is currently, they target Windows and Valve does the work if they care to make it work, just like IBM tried with OS/2 and its "Runs Windows better than Windows" campaign.
Also good luck supporting DirectX 12 Ultimate features not available on Vulkan, or what Microsoft and NVidia just announced at CES, AI Shaders.
If Nvidia and game developers care about supporting Linux they’ll try to figure something out.
If they don’t they won’t. But surely the cost required to do so is lower when targeting Wine/Proton than native Linux?
If they are unwilling to do that when the cost is relatively low (or at least e.g. making these new features toggle-able/optional) they certainly wouldn’t even provide a native Linux if Proton wasn’t a thing.
Exactly. If Valve went the native route why would that have changed?
> Which as IBM discovered
You keep repeating that in every comment but I don’t quite understand how is this some conclusive argument. Especially when you don’t elaborate on what specifically do you think could happen.
> If Valve went the native route why would that have changed?
We would have an ecosystem of native GNU/Linux games, and Microsoft could do whatever they felt like withouth any side effects.
> You keep repeating that in every comment but I don’t quite understand how is this some conclusive argument. Especially when you don’t elaborate on what specifically do you think could happen.
It will happen the same it did to DR-DOS, OS/2 and EEE PCs, and did to anyone else that built kigdoms in other people castles, the rest you can discover yourself.
Another example, even though Android is heavily dependent on Java ecosystem, and Google has won the lawsuit, they have started rewriting even the core components in Kotlin, because alongside JetBrains, they control the narrative of what goes on their kingdom.
Likewise Apple's approach with Game Porting Toolkit, is to help rewrite DirectX into Metal, not to run them unmodified, again control the narrative.
> We would have an ecosystem of native GNU/Linux games
Or we wouldn’t, perhaps a small fraction would be available on Linux severely limiting the appeal of devices running SteamOS etc. which would just give the entire handheld market to MS without them even having to do anything.
Of course it possibly wouldn’t exist in the first place because Valve would have had much less incentives to develop the Deck if the market for it would have been much smaller.
> Likewise Apple's approach
And nobody cares about it or gaming on Mac. IMHO despite all of Apple’s success gaming on Mac is in a worse state than it was back in the early 2000s.
My entire point is that however flawed Wine/Proton approach it is better than wasting money on a platform nobody cares about or develops for. I assume Valve could start paying developers to port their games to Linux or something but I can’t really think of much else..
> they have started rewriting even the core components in Kotlin
Exactly. Valve can start doing the same when they gain enough market share to justify this. If Linux gains at least 10-15% of the PC gaming market game developers won’t be able to afford to ignore it anymore. Anyway that’s a relatively nice problem to have compared to the alternative.
I hope steam machines make a comeback. Steamdeck is the best gaming platform I've ever owned, and I'd easily pay up to $2k for a box that I can just put under my 4k TV and have all my 1000+ steam games there without ever having to see the windows logo.
I likewise hope they do, but that said once you get into the >$700 range a custom PC build starts to be a much better value. With projects like Bazzite being IMHO even better than Steam OS and having full support for desktops, I suspect a higher end box would have a lot of competition from DIY (which is still pretty popular in the gaming world).
I'd love to see them dip a toe in with an ~$700 NUC-style steam machine with higher specs than the Deck but still very affordable for most people. If that sells well then increasingly look at more premium models, though it may be in their best interest to facilitate custom builds at that point. They could possibly even sell the parts or a kit or something!
They want you to be able to play (and buy) games from Steam. I doubt they're fussed about whether you're using their official image or a homebrew enthusiast build.
Bazzite now (as of last week) has beta support for nvidia cards. Only through the open driver though, so no series 10 cards unfortunately (I have one laying around which I was planning to use to build a gaming box on my living room).
You can pretty much get that experience today. Go build yourself a gaming PC and install Bazzite. It is a bit focused on handhelds but also works on more desktop form factor but controller focused PCs. Steam Big Picture works as well so it is a lot of the same interface as what you'd get on SteamOS.
It's funny that they still call it Big Picture Mode, because the UI makes assumptions about being on a Steam Deck (definitely in Linux, and even in Windows).
In fact, my Legion Go counts as a Steam Deck in my Steam Replay stats.
(I guess the actual funny bit is that they make assumptions that a general-purpose UI is only used on their hardware…)
I'm afraid that the times where you do have to interact with Windows are exactly what some people are trying to avoid — the nags to upgrade to Windows 14 (remember when 10 was the last one?), defender popups, ads for 3rd party bloatware when you connect a mouse, forced updates before shutdown, etc.
To be fair pretty much all of that stuff can be disabled by tweaking a few settings (on pro/enterprise at least). I do get the update notifications but that’s it.
90 comments
[ 4.4 ms ] story [ 212 ms ] threadAnd original Lenovo press link is https://news.lenovo.com/pressroom/press-releases/lenovo-legi...
I love the Steam Deck and they do a really, truly great job at tuning profiles of most games to run well, but the most demanding AAA games do sometimes hit a perf ceiling.
If you're curious about the details and you haven't seen it before, you can enable metrics/stats in settins and see a screen overlay of your performance. As someone who frequently keeps htop open on my laptop while working, I appreciate this feature :-)
> ...the world’s first officially licensed handheld powered by SteamOS™.
https://news.lenovo.com/pressroom/press-releases/lenovo-legi...
https://store.steampowered.com/news/app/593110/view/52983491...
Both in the context of the handhelds themselves (do they meet the steam deck's high bar for reliability and longevity) as well as how the companies handle things like open source contributions or vetting games.
It takes more than just the OS to replicate the steam deck's success (even though SteamOS itself is very well executed)
Wine is way more than a DirectX implementation.
I never even use my phone as a phone. I certainly don't make calls while I'm out and about. The thing pointlessly enables spying on me when I don't even want it. But a pocket computer with wifi, enough storage for my entire music collection + wikipedia + maps, and ability to take phone-quality pictures? Running an actual Linux distribution so I don't run into stupid issues like how Android doesn't let you use NFS so they can push you into their cloud services? Awesome.
It's quite a bit larger (and more powerful too).
At that point I didn't even own any Steam games and hadn't played a game in well over a decade, and had no interest in doing so. I'm a sucker for open hardware though. Forward a few years and I've been really happy with it. With games so easily accessible I have bought quite a few over the years now.
Adding cameras would be awesome! Though were I in charge of product for it, I'd want to think very carefully before doing so. It is a general purpose portable computer, but it's important to remember that the vast majority of the market doesn't buy it for that, and a camera and associated software may get in the way of it's primary purpose... Though, it could enable some pretty slick live video chat during multiplayer, which could be pretty awesome for people that play those kinds of games.
I've been using Jovian-NixOS as a placeholder for SteamOS, and have been disappointed at how mouse-and-keyboard-centric Linux UIs still are.
All it's really doing is pulling a LineageOS image, then dropping in the waydroid "drivers/firmware" that make it work with wayland using a series of patch files people have contributed. It all then gets run in a container which you "remotely" access for the GUI parts. Starting and stopping the container is like turning your phone on and off. If you set it up as a systemd service then you can just make this lockstep with your actual boot sequence.
It's primarily "marketed" as a tool to run android apps on Linux, which means the documentation kind of guides you down this path of running individual apps as their own window in Gnome or something. This doesn't work very well and apps tend to crash when you try to resize them. However, if you run it via `waydroid show-full-ui` it works perfectly in my experience and operates just like a phone/tablet.
I currently use it with a 2-in-1 x1 Thinkpad (meaning it has a touchscreen + pen), primarily for reading comics, books and watching movies when I connect it to my TV. I have it "wrapped" in Hyprland, so I can easily switch workspaces to return to the "host" if I want to, but I usually just do everything entirely within waydroid, and lazily use JuiceSSH for most host operations.
My laptops camera is a bit special and lacks linux support, but my understanding is a properly supported camera would work fine with waydroid. It doesn't support screen rotation or wifi configuration though, so those parts still need to be handled by the host system. I have my laptop auto rotate via hyprland since it was the easiest but I think it would be trivial to just add a patch for support as well.
It uses a pretty old LineageOS image, but I rebuilt it with some community patches so I could use the latest image, so you aren't even restricted to just LineageOS. I also didn't quite like how my stylus was being treated by android, so made a custom idc file for it and plopped it into the overlay filesystem waydroid has, saving me the effort of having to rebuild the image again.
It all sounds kind of complicated, but you can get most of this up and running right out of the box, complete with google play services too (requires you get a token from google, but they have guides for this). It's just once you notice something bugging you, you can actually do something about it, in an afternoon even, instead of just dealing with it forever. Something I personally appreciate.
That said, I'm using it as a tablet, for a phone there will be some gaps that need still need filling, such as:
- You would need to somehow get a small touchscreen device and install linux on it
- GPS: You would need to make sure this device exposes the GPS interface in some way and integrate it with waydroid. Since waydroid works via the same path OEMs take, there are probably some guides for this.
- LTE: Same as GPS, though so long as you can get internet on the host, waydroid will inherit that automatically.
- ARM Apps: Waydroid has some stuff that should allow you to run ARM apps but I've never bothered to set it up, not sure how well it works.
What should work out of the box, assuming they support linux:
- Cameras
- All touch interactions
- All stylus interactions
- Keyboard + Mouse
- HW rendering, works perfectly
- Network access (only configurable on host though)
- Direct USB access (out-of-the box in the sense you need to enabl...
See Librem or PinePhone/tablet for instance.
I've played half a dozen games on the original Go, and I don't think I've used more than one rear button. In Half Life, it's more comfortable to hold for crouch. In one of the Spider Man games, it expected a trackpad click, which was easier to perform with a rear button.
Since games have been designed to be playable with only the face buttons, there's not always obvious utility for extras. Maybe as premium controllers become more popular, this will change. For now, I can definitely see the argument that 4 buttons are superfluous (except for standardization).
Rollercoaster Tycoon, FTL, point and click games are games the trackpad has vastly improved for me.
Even for emulators, they're great for binding all kinds of extra functions.
Using the right trackpad also means I can free up the right joystick for other stuff, and there's always enough other stuff I want to access regularly enough.
This reads like "I never play mouse driven games"
Like how ds4windows is used to translate Playstation controllers gryo and touchpad. Even the Steam controller requires Steam. I don't think oems are investing in the software to do the same.
I think the Orange Pi Neo is the only (other) one that's targeting Linux only. And they do have touchpads
Lenovo's current handheld exposes an Xbox controller and a USB device with extra buttons. Handheld Daemon synthesizes them into a virtual PlayStaton controller. If you tried to use the current SteamOS without Handheld Daemon, the 6 extra buttons wouldn't work. The community is waiting to see if Valve ships a more robust solution that Linux can standardize on, or if we'll still be using daemons like HHD for devices that don't have official Valve support.
It is how i got to play the original System Shock 2 on my Steam Deck even though the game was designed with a keyboard and mouse in place.
That's why distros like Bazzite and ChimeraOS have an additional translator[0] to take input events and output a controller that Steam supports, Dualsense edge or Steam Deck's controller, to be able to add support for gyro and back buttons.
[0] https://github.com/hhd-dev/hhd or https://github.com/ShadowBlip/InputPlumber
This is something to watch : we'll see how much Valve actually cares about Linux (and libre software/hardware in general) compared to just using it to get rid of Windows : will they make their new controller API SteamOS exclusive ?
(How well do Steam Link and Steam Controller work without Steam ?)
Steam Link the streaming application does require the Steam. Steam Link the discontinued hardware was able to run local applications without Steam.
The Steam Controller defaults to a keyboard and mouse mode until it receives a signal from Steam. So even on Linux, the Steam controller required Steam to function as a controller until the protocol was reverse engineered. Can read about it in the driver comment header, and the Valve copyright that didn't get added until they contributed support for the Deck controller in these drivers : https://github.com/torvalds/linux/blob/master/drivers/hid/hi...
(I recently started again regularly using both the hardware Steam Link and the original Steam Controller - funnily enough, for a non-Steam game - I'll have to check out what people have hacked over the years...)
On the surface level it looks it bit like the half-assed Linux efforts for the Steam Machines and EEE PCs.
This time the software - if it works as well as on the Steam Deck - is arguably more user friendly than Windows.
So it would have been nice for linux to also be a "premium" option.
On the other hand the Go S with Steam OS has a real chance to compare well against its more expensive Windows version.
Lets be real, with SteamOS having a Windows ecosytem, it is about saving Windows licenses mostly, and not really fostering a GNU/Linux gaming ecosystem.
After I got everything set up (which wasn't hard), the amount of maintenance I've had to do on there is zero. SteamOS managed to make desktop Linux accessible to basically anyone; it was easy to build my own game console that I enjoy considerably more than my Xbox One that I never bother with.
Valve has made desktop Linux a lot more viable than I ever thought would happen.
You must have used NixOS before, or had hardware that worked perfectly with a default config.
If you mostly deal with AMD stuff, most things "just work" with Linux. In my case, WiFi worked out of the box, graphics driver loaded fine, sound driver worked fine.
In a market that's getting more crowded by the day, where so many of their competitors are using the same SoC and come in at roughly the same price point, it's a rational decision to control costs as much as possible. Not to mention, they also need to compete against the Deck itself, which Valve is subsidizing to hit its pricepoint. The alternative would be to innovate and differentiate your product offering from your competitors (like the original Legion Go, which had detachable controllers and a fancy pseudo-mouse). The fact that this machine didn't carry those features forward is a strong signal that the bottom of the market is where the early wins will be.
Keep in mind also that Microsoft is planning their own handheld xbox experience. Speculation: MS' effort will look a lot like Valve's, with Xbox games coming day-and-date to OEM-built handhelds, as well as a first party reference implementation that's intended as a benchmark for devs. I imagine that most OEMs are using the cheaper steamOS pricing as a hedge against future gouging from MS; if there's an "Xbox certification process" that proves to be too cumbersome, a vendor that's already shipping cheap(er) devices en masse with SteamOS has the option of simply dropping windows support altogether.
On one hand you have the shitty and/or DRM-plagued stores, on another you have DRM-free ones like GoG, Humble Bundle, Itch.io... which in my experience Just Work (tm) in a few clicks thanks to the likes of Lutris launcher/Proton/Wine.
(And in the middle, Steam(OS) to rule them all.)
P.S.: I've heard that Battle.net games also work easily and well, but I don't know how long this will last with Microsoft buying Activision-Blizzard and Valve ramping up their fight against Windows...
The day they decide it has been long enough, like they did with XP licence that ended up wiping EEE PC, Valve would have wished to have fostered a native GNU/Linux ecosystem instead.
> fostered a native GNU/Linux
Wouldn’t that require a stable ABI and such? Making Linux binary/proprietary software friendly seems like a much harder task than [potentially] fighting MS over the right to reimplement their APIs.
SteamOS can provide such ABI, just like Google does with Android.
Why build something from scratch when Wine/Proton is already there and is pretty much the same thing?
Of course that would be the trivial (if still very complicated and expensive on its own) part, incentivizing developers to invest the amount of resources needed to port their games to a new platform (which would have single digit % market share at best) would be a lot harder.
Sure there might be some risks. But I don’t really see why MS would be bothered to actually try and kill Wine. The risk of antitrust etc. would be very high and MS wouldn’t gain too much from this anyway.
Besides.. the Supreme Court already ruled in favor of Google over the Java/Oracle thing not sure how would this be much different.
Because Valve castle is on Windows kingdom, and it might learn about OS/2 the hard way.
And? I fail to see how is that so valuable on its own.
I mean don’t get me wrong I do see some merits but realistically if Valve went that way they would struggle getting of the ground and it would just be a waste of time/money.
Possibly one pathway that could actually be feasible would be to leverage Wine to gain significant marketshare and then incrementally transition to a more “native” approach.
> the hard way.
How? I mean what could MS actually do about this? Besides trying to compete by offering alternative options.
Also good luck supporting DirectX 12 Ultimate features not available on Vulkan, or what Microsoft and NVidia just announced at CES, AI Shaders.
If Nvidia and game developers care about supporting Linux they’ll try to figure something out.
If they don’t they won’t. But surely the cost required to do so is lower when targeting Wine/Proton than native Linux?
If they are unwilling to do that when the cost is relatively low (or at least e.g. making these new features toggle-able/optional) they certainly wouldn’t even provide a native Linux if Proton wasn’t a thing.
Porting is relatively easy for those studios, they don't because that development costs of having a Linux build and QA isn't worth it.
With Valve pushing WINE/Proton is even better, they don't need to care at all that Linux exists, Valve will do the work for free for them.
Which as IBM discovered, isn't really the future proofing way to build an ecosystem with someone's else OS.
Exactly. If Valve went the native route why would that have changed?
> Which as IBM discovered
You keep repeating that in every comment but I don’t quite understand how is this some conclusive argument. Especially when you don’t elaborate on what specifically do you think could happen.
We would have an ecosystem of native GNU/Linux games, and Microsoft could do whatever they felt like withouth any side effects.
> You keep repeating that in every comment but I don’t quite understand how is this some conclusive argument. Especially when you don’t elaborate on what specifically do you think could happen.
It will happen the same it did to DR-DOS, OS/2 and EEE PCs, and did to anyone else that built kigdoms in other people castles, the rest you can discover yourself.
Another example, even though Android is heavily dependent on Java ecosystem, and Google has won the lawsuit, they have started rewriting even the core components in Kotlin, because alongside JetBrains, they control the narrative of what goes on their kingdom.
Likewise Apple's approach with Game Porting Toolkit, is to help rewrite DirectX into Metal, not to run them unmodified, again control the narrative.
Or we wouldn’t, perhaps a small fraction would be available on Linux severely limiting the appeal of devices running SteamOS etc. which would just give the entire handheld market to MS without them even having to do anything.
Of course it possibly wouldn’t exist in the first place because Valve would have had much less incentives to develop the Deck if the market for it would have been much smaller.
> Likewise Apple's approach
And nobody cares about it or gaming on Mac. IMHO despite all of Apple’s success gaming on Mac is in a worse state than it was back in the early 2000s.
My entire point is that however flawed Wine/Proton approach it is better than wasting money on a platform nobody cares about or develops for. I assume Valve could start paying developers to port their games to Linux or something but I can’t really think of much else..
> they have started rewriting even the core components in Kotlin
Exactly. Valve can start doing the same when they gain enough market share to justify this. If Linux gains at least 10-15% of the PC gaming market game developers won’t be able to afford to ignore it anymore. Anyway that’s a relatively nice problem to have compared to the alternative.
I'd love to see them dip a toe in with an ~$700 NUC-style steam machine with higher specs than the Deck but still very affordable for most people. If that sells well then increasingly look at more premium models, though it may be in their best interest to facilitate custom builds at that point. They could possibly even sell the parts or a kit or something!
They want you to be able to play (and buy) games from Steam. I doubt they're fussed about whether you're using their official image or a homebrew enthusiast build.
You need a Radeon card though
https://bazzite.gg/
In fact, my Legion Go counts as a Steam Deck in my Steam Replay stats.
(I guess the actual funny bit is that they make assumptions that a general-purpose UI is only used on their hardware…)
Why? I mean you can pretty much just boot directly into Steam’s full screen mode and almost never interact with Windows already.
Too bad they couldn’t get a more up to date chipset in, but I suppose there will be a bunch of little tweaks to power management, compositing, etc.