Ryan Gordon and Ethan Lee on Proton and the Steam Deck

I'm coming to the conclusion that proton is inevitable.

"Write once, run everywhere"

Kinda goes straight into that philosophy and if I've heard of that mantra, I'm sure every developer trained has as well.

I wouldn't call devs lazy for using proton. Is it really that different from hitting "export" in unity in terms of results? Both fit that mantra don't they.

I can see the ideal here, write one game and know that a runtime/framework/interpretor/layer takes care of the nitty gritty so the game runs on everything. The OS doesn't matter.

I think I'd still prefer native. It seems more... Stable? But I wonder if that'll always be there case?

Quoting: MohandevirThe only part where I'm curious to know how it will be done, it's on the "Steam Deck desktop" side and all that productivity stuff... Will Valve create a SteamOS app store that integrates Proton too?

I think we might see more things like Blender on Steam.

I think this was inevitable, steam machines, link and controller made them learn a lot, and brought them to the investment onto d9vk, dxvk and wine/proton.

They did realize that they could not move a lot of developers to native ports because they did not reach a critical mass. They realized without the support they will not.

I think proton may be the only way because it's a compatibility layer they can support on their own.

While I could not care less if a game uses wine or not to run as long as it runs fine, I do personally not think that will be possible without a huge number of developers actually supporting proton/wine for the game developers. And just the caring about the bugs in proton costs money for a studio.

I do not see that, and while wine got a lot better in the years, there is so much which prevents a lot of games behaving properly.

Browser support, .NET support, sound, cutscenes, etc. are still lacking in areas, as is anti cheat, and while we heared about anti cheat, we didn't hear about any of the others.

A lot of games need Proton tweaks. Of course they could do that at valve, but that team would need to be huge, and often wold require a custom proton (and not just an own prefix) for a certain game to work around side effects which just do not occur in windows - since a best assumption how to handle certain things in linux compared to windows is just that, and there is a lot of that in wine. While 80 percent of the time accurate, it's not the other 20 percent.

Valve actively asking devs not to port doesn't sound great, though chances are it's because native ports (sometimes done in a hurry) don't work as well as proton, so it makes sense to tell devs to use proton if they are otherwise going to do a terrible job at porting. Valve probably has data here we don't have, so don't be so quick to judge.

Regarding native ports in the future, I think it's pretty obvious. If you are a dev and your game has 20m users on Linux, you *will* spend resources on a native port, simply because the risk of proton breaking your game and you getting 20m angry users is too high. So if the game is successful (and so is deck), it (or the next game from the developer) will get a native port.

Quoting: STiATBrowser support, .NET support, sound, cutscenes, etc. are still lacking in areas, as is anti cheat, and while we heared about anti cheat, we didn't hear about any of the others.

They addressed it in one of their FAQs and said that devs should use Vulkan in general, avoid .NET stuff and Media Foundation.

Quoting: dubigrasuwe have a saying: befriend the devil until you cross the bridge.

we also have a saying: the road to Hell is paved with good intentions.

Quoting: Guest

Quoting: Lachu

Quoting: GuestI don't see it good for GNU/Linux however. It's consolidating more gaming under the control of someone who is not the user.

But what else than Valve/Steam? Valve controlling large spectrum on gaming comercial-linux-market. I think open source are not so great than closed source/commercial. I do not tell open source software is bad. It is great. I use only open source software (excluding firmware, DRM for video watching, BIO, etc.), but no games. I am happy with that. If open source games give any alternative, I will use open source games.

I can paid for open source software. I was give donation to KDE team recently. Of course, I give a lot more for closed source software/games, but as regular user, I think open source software should give this benefit (cheaper), because it is a democracy/free market and in this case money are better used (I paid for what I want - for example - I do not like option A, I do not paid for it and it never will be added to some open source software, etc. ). Maybe that means, open source software developers get smaller amount of money, but I am not 100% sure. Firstly, there is no piracy and people will paid as much as they can. Secondly OS developers can collaborate with other projects and charity people (sorry, English is not my native) easier, so there is not the same cost of creating software.

In summarize, I can paid for Open Source software, but if Open Source games are worth it? Of course - freedom is very important and If we paid corporation, it will lobby to take of more of our laws, etc. But gaming is like watching movies - it is like communing with culture. If I do not commune with culture, I will be really freedom?

That's a really valid point. My own take on it is that I'd like to be as extreme as RMS, but in all practical scenarios it's not feasible from a cultural participation perspective (in this case, games). So I instead try for open source on the OS, open source to choose something to run; running a game is my choice, but what is needed to run the game should be open source.

The thing with trying to be as extreme as RMS, is there really isn't any hardware open enough to meet that standard. Would be an awesome, yet different world if there were!

Quoting: Lofty

Quoting: dubigrasuwe have a saying: befriend the devil until you cross the bridge.

we also have a saying: the road to Hell is paved with good intentions.

And the road to diabetes is paved with good chocolates... hmmm, chocolate...

Quoting: kuhpunkt

Quoting: STiATBrowser support, .NET support, sound, cutscenes, etc. are still lacking in areas, as is anti cheat, and while we heared about anti cheat, we didn't hear about any of the others.

They addressed it in one of their FAQs and said that devs should use Vulkan in general, avoid .NET stuff and Media Foundation.

It was this one.

Quoting: CatKiller

Quoting: kuhpunkt

Quoting: STiATBrowser support, .NET support, sound, cutscenes, etc. are still lacking in areas, as is anti cheat, and while we heared about anti cheat, we didn't hear about any of the others.

They addressed it in one of their FAQs and said that devs should use Vulkan in general, avoid .NET stuff and Media Foundation.

It was this one.

Which are quite good advises that will in most cases make the games run at very very good quality and speed. Chances are a developer exchanging .net and media-foundation stuff or implementing a vulkan renderer or doing whatever needed to get their game work well in proton will do that directly in their normal sources, so further support should be more or less a given. That indeed makes a big difference to the poor porting nightmare we regularly saw in the past. (Again, taking out Ethan and Ryan. They somehow managed to support the games they ported over time, while other porters didn't).

The final goal isn't every game getting ported to Linux, it's every new game project taking linux into consideration and using the tools that let them support us. Wine/Proton aren't without alternatives and weaknesses, yet they are a part of our (open source) ecosystem, without any doubt.


Last edited by const on 21 July 2021 at 6:44 pm UTC

Quoting: constThe final goal isn't every game getting ported to Linux, it's every new game project taking linux into consideration and using the tools that let them support us.

What I'd like to see is developers not thinking in terms of "ports" at all, but to have Linux builds as part of their standard development and testing routine for their PC release. They'll squash more bugs before release that way, just from having more lenses to examine their code with. They'll have saved time and money even without a single sale. Then they only need to make a depot of their existing, working, tested, Linux build the same as their Windows build.

I knew this topic would appear here when I saw on Reddit

First I understand Ethan's point, he's probably feeling "betrayed" because Valve's decision affects directly his revenue, but I think stop programming for good it's a little over reacting in this case (unless of course he was already thinking about that before and wants to do something else)

But I think Valve has a point in this case, they can't afford another failing like Steam Machines, which was a scenario where they depended entirely on third part developers porting to linux, and in the end we all saw what happened. Failing again and Valve may considerably reduce its investment on Linux gaming, and we don't have any other company to back it up in this case.

The best plausible scenario I can think is: People start to play using proton and developers can give some "support" to make sure it's working, then average users start to take more interest on "this linux thing" since they can play their games as they do on other platforms. After some time, some developers (now backed by some real mainstream base of linux players) start to think "hey, my game could perform better if I don't use this wine thing" and take interest in making native ports, after all, they'll have a public to justify it and the competitors may be already thinking on this too.

May this hypothetical scenario partially come true, and Top developers like Ethan may start to receive some requests from big companies too.

And for those saying Valve doesn't care about Linux: You're right.

I never saw Gabe Newell or any official announcement from Valve saying they are aiming on increase Linux adoption like some people seems to pretend that the case, after all most of their revenue come from windows players, Linux for them is only so they can have more independence and to make sure Microsoft won't interfere in Valve's business. And this is perfectly ok!

Valve has no obligation to us Linux users, we just have a seller/consumer relation like anyone else (they're not the "saviors" of Linux desktop) but fortunately their business models generates some benefits for us despite not being their main goal.

At lets be honest, a lot of us don't care about something like steam deck either. In my case for example I can't buy one even if I want, so my interest in this project it's the possible benefits of this project helping games to run better on my pc/notebook with [INSERT PREFERRED DISTRO], and I suspect this is the case with more people than they'll admit .

And this is perfectly ok too!

It strikes me that Lee's reaction is kind of forgetting that all the existing Linux desktops, that up to now have been the only reason for releasing native Linux games, will still be there after the Steam Deck releases. Even if people targeting the Steam Deck ignore native releases, that doesn't actually shrink the incentive to release native. So I think he might be overreacting.

Quoting: Lofty

Quoting: dubigrasuwe have a saying: befriend the devil until you cross the bridge.

we also have a saying: the road to Hell is paved with good intentions.

I've always hated that particular saying. Cuz, you know, it isn't. Sure, you can cherry pick cases where good intentions lead to bad outcomes. But come on, what's it supposed to mean? Does it mean you should have BAD intentions so you go to Heaven? What?
There's plenty of intentions out there that aren't good, and they are what cause most of the problems.

This blog post appeared on the GoL discord server:
https://mdiluz.micro.blog/2021/07/19/native-linux-ports.html
It's an interesting read and adds another perspective to that discussion.


Last edited by jens on 21 July 2021 at 9:24 pm UTC

Quoting: jensThis blog post appeared on the GoL discord server:
https://mdiluz.micro.blog/2021/07/19/native-linux-ports.html
It's an interesting read and adds another perspective to that discussion.

It is an interesting read. On one hand, it's pretty explicitly saying that Proton or something like it (with the needed ingredients of Vulkan and stuff like DXVK) is the solution.
On the other, it strikes me as implicitly (although the poster himself maybe has a blind spot for that because his background is very specifically in porting existing games) supporting the point made often but in this thread mainly by CatKiller, that it's best to develop cross-platform from the start rather than developing with an orientation to one platform and then having to make tons of changes to port to another.

Sad to here this from Ethan. This is called precarious employment. These contracts are favourable for Valve. It's a hire and fire philosophy. Even without having to fire anyone. I don't think it's okay.

Everyone needs a certain amount of planning security. You have to be able to look around for something new without worrying. I also don't know if Linux is really so far that you should let all the know-how go. How many people are there who know so much about Ethan's field? Sound problems are the nastiest in gaming. Something the brain can't block out. Makes a game basically unplayable.

Quoting: Purple Library GuyIt strikes me that Lee's reaction is kind of forgetting that all the existing Linux desktops, that up to now have been the only reason for releasing native Linux games, will still be there after the Steam Deck releases. Even if people targeting the Steam Deck ignore native releases, that doesn't actually shrink the incentive to release native. So I think he might be overreacting.

Well it's not easy to see all that you have spent thousands of hours building up come crumbling down, and having once been in the same situation I fully understand Ethan:s feelings here. I do hope that he realised soon however that he is one hell of a developer and that there are millions of other things than conversions that can use a man of his talents, then at some point in time conversions will be back again and he can return to what he loves to do.

After musing a bit longer, I think Ethan Lee has a point, but it's a limited one. That is, I think he may be right that it's going to be hard to get work porting games to Linux, which is hard on him and I'm sorry for it, and yet I think going forward a successful Steam Deck will still lead to more native Linux games.
Let me lay out a couple of distinctions. So, up to now, most games have run on Proton as it were by co-incidence. The makers of those games have no interest in Linux, may have barely heard of Linux, and the fact that the games run is all down to the hard work and talents of the Wine and Proton and DXVK developers. There seem to be a few exceptions to this, we've lately seen game developers say things about trying to ensure their game runs well on Proton, but it seems fairly rare. So, we have a big population of games and developers that don't give a hoot about Linux, and many of the games run on Proton despite this.

So then there are the somewhat smaller group of developers who do care about Linux somewhat. What impact has Proton had, up to now, on that smaller group? Well, it doesn't seem to have really stopped people from developing games in cross-platform ways that include Linux, as far as I can tell. There's still a strong stream of new games supporting Linux from the beginning. Maybe it's tailed off a bit since the heady days when people thought the Steam Machines might be a thing, but that's been a long gradual thing and I haven't noticed it getting worse since Proton. One might have expected that to happen, but I don't think it did.

But it does seem like the existence of Proton has already reduced the viability of the porting business as such; note the way Feral has basically moved on from doing that and nobody has really replaced them. Why spend all that money and effort porting an existing game when it probably works fine on Proton already, or can be made to do so pretty easily?

If the Steam Deck succeeds, it seems like it will greatly increase the number of developers giving Linux a thought at all. Steam's increased emphasis on Proton, both in terms of telling developers about it and improving the technology itself, does make it even less likely that developers thinking about Linux will bother porting existing games. And even though there will be a lot more of them, Ethan Lee might be quite right that the chance they'll bother doing a port will drop so low that there will still be a lot fewer ports happening. Anyone who just started thinking about Linux because of the Steam Deck will, when looking at their back catalogue, surely conclude that as long as the games run on Proton that should be fine--any improvement wouldn't be worth the cost and effort of making a port. Ryan Gordon may indeed be disappointed if he tries to persuade more people to port their existing games.

But new games, and the decision to target Linux or not from the beginning, is a different question. Going forward, there will be two factors--on one hand, Linux has far more visibility and users than before; on the other, that target can at least somewhat be satisfied by paying attention to having it work in Proton. Which factor will dominate? Based on what we've seen to date, with that side being less impacted by Proton, I think it's likely the first factor will dominate and overall, while some developers may not build for Linux because Proton is good enough, the gain from it being a bigger platform will be greater.

So if I had to make a guess I'd expect, if the Steam Deck is a big success, in the end we'll see fewer ports but more games built Linux native from the start. Which suggests that Mr. Lee and Mr. Gordon's best bet might be to shift to consulting on how to properly do cross-platform Linux friendly development on new games, rather than porting old ones. Although I'd certainly be happy enough to see Valve hire Ethan Lee, as he suggests, to work on infrastructure.


Last edited by Purple Library Guy on 21 July 2021 at 10:44 pm UTC