Is Valve Working on a Windows Compatibility Layer?

Ryan Gordon says that Vulkan is going to be a better alternative

This is the latest rumor that’s been hitting the interwebs! Apparently this speculation is coming from Ars Technica, which is not your average tabloid. There’s not much substance per se to report at this stage, but some elements are indeed intriguing, especially the existence of a “Steam Play” mode,  along with the following properties as reported by SteamTracking:

  • “Steam_Settings_Compat”
  • “Steam_Settings_Compat_Info”: “Steam Play will automatically install compatibility tools that allow you to play games from your library that were built for other operating systems.”
  • “Steam_Settings_Compat_Enable”: “Enable Steam Play for supported titles
  • “Steam_Settings_Compat_Forced_Info”: “You can use Steam Play to test games in your library that have not been verified with a supported compatibility tool.”
  • “Steam_Settings_Compat_Forced_Enable”: “Enable Steam Play for all titles
  • “Steam_Settings_Compat_Advanced_Info”: “You may select a compatibility tool to use with games that have not been tested or verified to work on this platform. This may not work as expected, and can cause issues with your games, including crashes and breaking save games.”

What can I say? This looks like the real thing! If it is really happening, I cannot say it would be a huge surprise either: adding WINE support directly thru Steam has been a request from Linux gamers (at least not the “native” zealots type) for about forever, and from a feasibility standpoint we have been getting closer and closer to it every single day, with WINE and recent initiatives like DXVK too. I also suspect Dotsujin who is behind the DXVK project may be secretly working for Valve (at least as a contractor), since he systematically closed any topic related to sponsoring via Patreon and the like. Of course, there’s always the possibility that he’s a lone coder with a lot of time on his hands, but that sounds a little unlikely, given how fast he is advancing on this project.

Now, do I believe in this pre-announcement/leak? Yes. I think this was the play for a while already. I cannot say anything more but this seems consistent with what I have heard thru other channels, while unofficial. Remember also what Pierre-Loup Griffais from Valve said just a few months back: 

Pierre-Loup Griffais: We also have other Linux initiatives in the pipe that we’re not quite ready to talk about yet; SteamOS will continue to be our medium to deliver these improvements to our customers, and we think they will ultimately benefit the Linux ecosystem at large.

This will have major implications for Linux gaming, no matter how you look at it.

First, this should help bump up the amount of games one can play on Linux/SteamOS from one day to the next. Of course, that should be true for indie games but more importantly for AAA titles which are critically missing in numbers right now. This means that anyone who was considering changing ships may suddenly see this as a more realistic endeavor than ever before. For Linux gamers who are solely playing games on Linux, this is almost a dream come true.

Second, while a lot of Linux gamers relied on Lutris right now for their gaming needs, having Valve involved as an intermediate should certainly mean something in terms of support/testing (at least let’s hope so…). The above descriptions indicate that some games will be “enabled” by default while others could be enabled “at one’s own risk”. This is a great approach to experiment and give feedback as “Steam Play” hopefully gets better over time.

Third, the impact on the porting scene may be huge. We have specialized companies making individual ports right now, who may be deprived a good reason to do so in the future as Steam Play rolls out and is “good enough” for all practical purposes. Would you prefer waiting a year to play the latest AAA titles with good framerate or playing it now, at release, with 10-20% performance loss? For me the answer is pretty obvious. So, what will happen to Feral, VP and the like? Will they still have a strong reason and financial incentive to produce ports, especially if they don’t get released on Day 1?

Fourth, this may be the first good opportunity to launch the Steam Machines again. And guess what? I would expect them to be AMD-based rather than nVidia since Valve is directly supporting the development of Mesa internally. Now, such Steam Machines would have a better shot at being more than nothing, if they support a very large catalog of games. At least, nothing like the no-launch from a few years back. And who knows, this time around Valve may turn around and reveal some actual custom-made hardware instead of relying on third parties. At least that is what I hope they would do.

Finally, I have read here and there that this is a “terrible” piece of news for Linux gaming. I completely disagree with such opinions. Having to worry about ports is what is truly terrible. Having to wonder if a promised game is actually coming or not is a shitty place to be in. Having to wait 6 months, 1 year, or 2 years to play the game that all your friends have already finished, is not good if you like to be involved in communities. The fact that you may not have to think anymore about “does this work on my OS?” but just start any game that you want to play is really what constitutes a console-like experience. And my guess is that our “ports” will end up being better this way – a Windows game is likely going to be more polished and bug-free than a rushed port, or a port that’s improperly tested on many Linux distributions. I have seen enough shitty Linux ports up till now to be convinced that we aren’t making good progress on the quality front. And good ports, like Feral’s, require lots of efforts and thus, time. We are not going to get past a few AAA ports per year this way, it does not scale.

So, an approach akin to emulation is ideal. It is likely to solve many problems, and performance-wise I can say we have a pretty good proof of concept with DXVK that we can hit high frame-rates even for DX11 games. And DXVK has not even been around for so long yet (and they have just added DirectX 10 compatibility in last week, making it possible to run Crysis).

In the end, this project is not for “us” Linux gamers – which is why some Linux gamers complaining about this are mistaken. It is clearly for Valve to build their own platform based on Linux, not tied to Microsoft. From there on they can have a very solid foundation to deliver on hardware projects.

For us, there is nothing to lose, quite the contrary.

BoilingSteam lets you access our content for free, but writing articles is a constant investment. We don't use ads or sponsporship, help us make our activities sustainable by donating via Patreon or LiberaPay if you prefer it anonymous. You can follow what we do via our newsletter, our RSS feed, our Mastodon profile or our Twitter feed. We also have Peertube, Youtube and LBRY channels. If you'd like to chat, you can also find us on (what is Matrix?)

0 0 votes
Article Rating
Notify of

Oldest Most Voted
Inline Feedbacks
View all comments

Then Carmack was right after all?


This speculation is such great news. For so long it seems like Linux was getting close but now we really are!