"Linux-native" is an outdated mindset. With tools like WINE and Proton+DXVK/D3DVK / SteamPlay, software developed on Windows is in all practicality platform agnostic.
After more than a decade, we know the cost of supporting Linux native builds seems to be the sanity and good will of the devs. For indie and small teams, it's literally taking food out of their mouths asking them to use resources fixing obscure bugs from a vocal minority that they could instead put towards developing new content or planning their next projects.
I get the sentiment, but really, we won the lock-in war. It's kinder to ask devs to stick to what they know and are comfortable with, and if they want to build with SteamPlay in mind, then we all win.
I used to care about this but then I just started enjoying my games. Almost every single player game I play in proton works so flawless you would have no idea proton is involved. While many of my linux native games fail to launch because they are dynamically linked to some old package which doesn't exist anymore.
Really? I find that I can play most games just fine under Linux, except those that require annoying anti-cheat software (but then I wouldn't want to run that under windows either...).
For my library, it's still largely a 50/50. But it has been getting better lately. Maybe it's 60/40. A big one is Space Engineers, which is supposed to work with such and so patch, but not for me.
While this is true, some devs in the indie community develop on linux[1] so support is there by default. It's also a small(hopefully growing) niche to target Linux as a platform. SteamOS is also a thing, albeit without a solid market afaik.
SteamOS isn't really a thing anymore. There's been no update in about 2 years. It's possible Valve is still putting some effort there, but hope for its future is pretty much just speculation at this point even if Valve hasn't abandoned it completely.
I feel like a good chunk of that is market politics: Valve has an interest in Linux support so they can (re-)build something like SteamOS. And they want that ability so that Microsoft doesn't get too stupid ideas of forcing restrictions on Windows gaming/tying it to the Windows Store/... But the obvious option of that is enough, they didn't need to fully push the SteamOS/SteamBox concept after it fizzled out, the fact that they could is cheaper and serves the purpose.
That is a very interesting view on things, and I would not be at all surprised if it's the truth. Valve need only devote minimal resources to keep it shelved while making sure it's updated just enough to be able to rapidly deploy, at the same time that ProtonDB ensures a good % of the library is immediately available.
Gamers are minority of PC users, but a little over 10% of PC's shipped are still gaming-level. Some of those are probably purchased for graphic designs and/or video editing, but the gaming segment would still be substantial enough that Windows wouldn't want to lose it. Not because of the immediate revenue loss: In the beginning you'd still probably have to pay a MS tax if you weren't building your own PC, or you wanted a laptop (and gaming laptops are a large part of the gaming market).
Once 10% of the market was installing Linux, that's a foot in the door MS really wouldn't like.
After more than a decade, we know the cost of supporting Linux native builds seems to be the sanity and good will of the devs. For indie and small teams, it's literally taking food out of their mouths asking them to use resources fixing obscure bugs from a vocal minority that they could instead put towards developing new content or planning their next projects.
I get the sentiment, but really, we won the lock-in war. It's kinder to ask devs to stick to what they know and are comfortable with, and if they want to build with SteamPlay in mind, then we all win.