Steam's top releases of May show why Steam Play is needed for Linux

Quoting: Linuxwarperthere is also the risk of Proton not working anymore. When devs develop for Linux they will experience issues, file bugs and then get it all fixed.

Not to discount the rest of your argument, but I don't find this part to be particularly tied to Proton. We've seen native ports (and software in general) become broken and never get fixed, and Proton support doesn't have to be a one-and-done thing: they can keep supporting and patching the game to maintain Proton compatibility. It's more a matter of how committed a dev is to supporting Linux than how they choose to do it.

Quoting: BeamboomDon't waste time on the less profitable markets unless there's nothing more to exploit in your current market and you need to focus on secondary markets for further growth - and by then you're probably better off defending your current position in the main market.

Not when you are aiming to reach more users, instead of "I want moar $$$". It's not about greed, but about creators increasing their audience. This topic really is about the commercialization of art in general. If you view art as pure business, that art quickly becomes garbage.


Last edited by Shmerl on 2 July 2019 at 4:48 pm UTC

Quoting: Beamboom

Quoting: LinuxwarperWhen and if that time ever comes, "No Tux No Buck" will be as meaningful response as it ever can be.

... But why? Why does it matter, if Proton provides a 100% working out of the box experience on par with a native build?
If you experience no difference?

Mind you, that is a rather big if. "Experience very little difference" I'll buy. But in theory, all else being equal, Proton has to lose a bit of efficiency compared to a native version, not so? And targets will move--not so much existing games, which will probably keep working on Steamplay at least as well as native Linux games. But new games will keep on doing new things, relying on new Windows-only middleware that it will take time for Proton to get working. So for a Linux relying on Proton, there will always be the risk of delays--or even complete failure to get something running--even if we gain market share that would otherwise guarantee same-day releases.
So as long as we're relying on Proton, Linux will be a second class gaming OS.

And then there is, as Linuxwarper mentioned, the ecosystem. If people are making Linux games, they are using libraries and middleware that run on Linux, probably cross-platform, so such libraries will gain prominence/marketshare. If those libraries and middleware running on Linux are open source, they may be helping to improve them. If people are making Windows games, they are likely using libraries and middleware that are not cross-platform. And these libraries and middleware may well represent more of a "moving target" than Windows itself; near as I can figure half the problem with Wine has never been Windows itself, it's all the dang .dlls.

So yeah, I think it matters. Proton will not still be a solution for a Linux that gains a stack of market share. It's a convenience today, and hopefully a bridge to a better situation tomorrow, and it seems to be very well written and I have tons of respect for the people who wrote the code, but it's not the ideal.

In the meantime, We loose marketshare...

Quoting: LinuxwarperBecause them developing their games properly for Linux will help the ecosystem

Yeah, ok, you put this in the context of the broader picture. I do agree with you from that perspective - it will undobtly benefit the overall ecosystem of Linux.

Quoting: ShmerlNot when you are aiming to reach more users, instead of "I want moar $$$". It's not about greed, but about creators increasing their audience. This topic really is about the commercialization of art in general. If you view art as pure business, that art quickly becomes garbage.

Gaming is business. A major player in the entertainment industry, an industry with millions of artists doing fantastic work, from musicians to actors, 3D artists, makeup artists, game designers, composers, painters, map designers, you name it. Art and business can easily go hand in hand, in many cases great art is also good business.

The gaming market is massive. No indie has tapped all the potential on their current platform(s). Nobody needs to support Linux in order to reach more gamers. They need to be particularly interested specifically in our segment. The rest have millions to struggle with reaching out to on just that one platform. Or branch out to the consoles, where a new massive market awaits.

And it's not about greed. It's about making sensible decisions. Even the most credible artist want to survive and live on their art. And to survive you need to run a healthy business. Ideology is what you can afford when your business is doing good.

Quoting: Purple Library GuyMind you, that is a rather big if.

Absolutely. Indeed it is. As experienced myself, with the current generation of Proton.
But - IF - it works out of the box (and indeed many titles already do), I really do not mind if I lose a few frames a second. If that's all I have to sacrifice from not running a native binary.
IF.

Add to this the sad(?) fact that indeed Proton have been observed to perform better than the ported games (compared to the Windows performance), and this picture is really starting to get a bit convoluted - and really quite interesting.


Last edited by Beamboom on 2 July 2019 at 7:08 pm UTC

Quoting: Purple Library GuyAnd targets will move--not so much existing games, which will probably keep working on Steamplay at least as well as native Linux games. But new games will keep on doing new things, relying on new Windows-only middleware that it will take time for Proton to get working. So for a Linux relying on Proton, there will always be the risk of delays--or even complete failure to get something running--even if we gain market share that would otherwise guarantee same-day releases.

Actually, in my view, this is the point where our (hypothetical) increased market share starts to work in our favour. Devs will be used to the income from Linux gamers making up a substantial share, thanks to Proton. They'll have a clear choice between building software with cross-compatibility in mind and cashing in on all sales at release, or spending additional time (and wasted effort) after release to make their game playable on Linux and grab the rest of the pie. From a practical standpoint, and again this depends on Linux being a sizable market, the choice to rely on cross-platform middleware from the start becomes basically common sense (well, not that it isn't already, but it becomes more obvious to people with dollar bills in their eyes). Companies that act only on profit not only want as much money as possible, they want it as fast as possible.

But either way, we're not even close to that yet. For now we need to continue our slow growth and rely on motivated and like-minded developers who care more about getting their passion projects into the hands of more people while making a decent living than about squeezing more and more money out of our pockets. Fortunately, those do still exist :)

Quoting: SalvatosNot to discount the rest of your argument, but I don't find this part to be particularly tied to Proton. We've seen native ports (and software in general) become broken and never get fixed, and Proton support doesn't have to be a one-and-done thing: they can keep supporting and patching the game to maintain Proton compatibility. It's more a matter of how committed a dev is to supporting Linux than how they choose to do it.

Quoting: BeamboomAdd to this the sad(?) fact that indeed Proton have been observed to perform better than the ported games (compared to the Windows performance), and this picture is really starting to get a bit convoluted - and really quite interesting.

Any software that is abandoned will sooner or later break as things change. It's nothing special to Linux as you all know. But what's to take note is that on Linux this happen more. Developers make a port, release it then they abandon or neglect the game. There are other factors that make it more likely that "native"/native ports will have higher risk of issues; developers have less experience with Linux, they have used software components that can cause issues (D3D, DX sound, their own anti cheat solution etc), sloppy development and so on.

Deep Silver supporting Steam Machines is what spurred the Metro Last Light port. Steam Machines didn't materialize to be the big thing they hoped for, so they have abandoned the port. When you go to adjust graphic settings all you have is a quality slider, and that's not only issue with that "native" port.

Native games underperforming or breaking is because of developers and not inherently because developing games on Linux is bad. Even Feral can't do anything about this, with Tomb Raider underperforming compared to Windows. If Feral was on the task to make a Linux version, while the game was being developed, you can bet the performance would be comparable to Windows.


Last edited by Linuxwarper on 2 July 2019 at 7:36 pm UTC

Quoting: GuestYou've pretty much hit the nail on the head here. There's no reason a native GNU/Linux game can't perform as well as a Windows one, but porting after the fact with a complex codebase that's designed around another platform, that's going to be troublesome no matter what (and is why there are so many crappy console to desktop ports for Windows).

Doom 2016 runs so well through wine because it's very similar to just using some cross-platform screen creation and input handling. The entire engine is designed around Vulkan, but I also suspect designed around a POSIX environment. Because of the server side, they had full support for case-sensitive filesystems in the core of the engine. All their memory management is perfectly compatible with anything - again, they need it to be because of server side. So in a very real sense, it was development with a native *nix environment in mind, and the end result is a flawless running game.

If you can get in while the game is being developed, and support GNU/Linux at that stage, then it becomes a much better end product compared to porting after the fact.

(Kinda hoping Stadia will help this aspect of gaming - even if there's no official native port, the design of games might progress to a stage where they're much better aligned with GNU/Linux environments.)

This whole ordeal with native vs wine is precisely why I am convinced Proton is a benefit more than anything else. Yes, it can and will make developers decide against a native port because it runs with Proton but that I think is for better. Some developers are in dire situations and may not be able to justify a port (costs to much), and so a Proton route would work for them. While others may be lazy, but if that's their mindset then it's likely that they would have developed Linux release poorly. Just to get it out of the way and check that "Linux Kickstarter" promise.

I am confident Stadia will help Proton question is just how much. As we already know the game developers will need to use Vulkan. So with that in mind these are the noteworthy games:
Baldursgate 3 - Previous games have had Linux release
Borderlands 3 - Previous games have had Linux release
Destiny 2 - Releases on Steam. VAC or/and Steam trustmatchmaking for anti cheat?
Doom Eternal - DRM free or will Denuvo version work with Linux?
Metro Exodus - Feral magic
Shadow of.. - Feral magic highly likely

More games will come from other developers as well.


Last edited by Linuxwarper on 2 July 2019 at 9:25 pm UTC

Just going to leave this here, platinum rated game compatible with Proton just released a native Linux port of their game.

Clearly Proton existing didn't kill that native port. In fact, likely because the game had a platinum rating, it probably got a stack of Linux gamers as a result, Linux gamers who looked at ProtonDB and said "I'll buy anything platinum since that seems like a safe bet'. The devs would have seen those stats before doing their Linux port.

Quoting: LinuxwarperNative games underperforming or breaking is because of developers and not inherently because developing games on Linux is bad.

Absolutely. There's many examples (outside gaming) where the native Linux version outperforms the native Windows version. No reason why it shouldn't be the same with games. If - IF - they are truly native code. But:

Quoting: LinuxwarperEven Feral can't do anything about this, with Tomb Raider underperforming compared to Windows. If Feral was on the task to make a Linux version, while the game was being developed, you can bet the performance would be comparable to Windows.

... The uncomfortable reality that too many miss in this, is that from a technical point of view there's not really that much of a difference between the AAA "ports" and running a game in Proton. Feral uses their own set of libraries to translate APIs as well - that's how they could release such big games relatively fast after a release, and with an as stable performance as they have (very few Linux related issues). And this is the very reason why we get an overhead and loss of performance compared to Windows. It's unavoidable.

Heck, people even called Virtual Programming releases for "ports" - their work is practically Wine bottles in closed source software.

So in may ways, if we call their work ports we might just as well call what Proton does as a "generic porting tool" of Windows games.

And it is also worth noticing that the work on Proton does indeed contribute to the Linux ecosystem - to my understanding more than Feral has, to pitch those two against each other.

So the reality here really is not black/white at all.


Last edited by Beamboom on 3 July 2019 at 9:56 am UTC