So, at least from personal experience yes, it makes absolutely no sense to port to Linux, if the result is still worse than proton.
Regularly test under proton during development, and you got a Linux Version of your game for free. From my experience, epic already took good care to make sure unreal engine works well under proton, so if you just pick that and don't add a ton of additional stuff that uses esoteric winapi stuff, you're already set.
We all know how that ended.
Another comparison that I think is more apt: "IBM PC Compatibles run IBM PC software better"
(In that sense, developers (or Linux porting studios) switching from their in-house compatibility layers to Proton isn't really a change in terms of those ports being 'truly native' or whatever.)
I wonder if what we're starting to see, rather than the inferiority of source ports, is actually that some of these alternative compatibility layers are just no longer keeping up with Proton, since Proton/WINE has such wide use/testing and rapid development these days.
My fucking linux machine runs on intel hardware that is targeting an arch made by AMD (x86_64).
There is no problem with sharing stable interfaces. The idea that we should not treat these as linux games because they consume an interface developed by MS, that is implemented just fine by proton on linux is... very very backwards.
Who gives a flying fuck who made the interface, as it can be implemented across platforms, I'm a-ok with the result.
You don't have a compelling argument here - step back, pause, and consider whether you're having an emotional reaction.
Your argument is basically "Developers should do things I want, for my benefit, with no regards to the success or consequences of those actions".
And that's a pretty fucked up argument.
A much better argument is: Proton makes linux a compelling platform for gaming. There is a decent chunk of market segment that keeps Windows around only for gaming. Allowing those users to move entirely to linux increases the appeal of developing native gnu/linux games, because there is a much larger native market.
This is how interfaces work - they allow you to develop new products that are compatible with the existing userbase. It's why they're incredibly important as stepping stones.
---
Side note - Proton came out of codeweavers, codeweavers develop a custom version of Wine. Do you happen to know what Wine fucking stands for? hint:
Wine Is Not an Emulator.
These days they check both builds if available and suggest the devs which one to use by default; often the Windows builds over Proton are simply better than mediocre Linux support.
Shipping something on Android does not guarantee functional builds for regular Linux, and PlayStation uses a heavily customized FreeBSD fork. Neither of which are helpful for much of anything.
Android even more so, assuming those that rely only on AGK/NDK.
1. Make wine really really good.
2. Start adding Linux-only features.
hehehe
If Valve gets gaming studios to target proton and guarantee compatibility, is a win.
If that's equivalent to gaming studios explicitly targeting MAME and QAing it, I'll take it.
Anyway it hardly matches Switch sales, for which studios do actually bother to actively support.
The magical thing is that they've done so much work with Proton that a device like the Steam Deck is more than a curiosity. It is a viable product. It is viable to play games on Linux in a way it has never been before. Even better, they aren't keeping it to themselves.
A purist ideology about linux isn't going to change reality. But Proton is. The Deck is a wildly popular device and the knock-on effect is going to be more systems running linux for regular people. The games may not be native but if they're indistinguishable from native games, then who cares? Meanwhile adoption of linux as a viable platform still goes up. Maybe on a long enough timeline things swing around, though I still very much doubt that.
The Year of Linux on the Desktop is still a fantasy, but reading the steam hardware survey, it is more true than it has ever been before, thanks to the Deck and Proton.
Idk what the transition plan could look like, but eventually we should hope to see the Steam Deck treated like any other console with highly optimized, native ports, and then hopefully some of that can be reused for 'more native' general Linux ports.
Not to say the current implementation is secure though, only that the use of Wine/Proton makes no difference.