Fair. I’ll take a working Windows build with proton over a janky Linux port any day of the week
As a cross platform developer I consider this incompetence.
That’s not necessary a bad thing. The world is full of less experienced programmers. But they’re making it look like it’s a hassle to release for Linux when in reality you can foresee and plan for this from the start, without much overhead down the line.
I completely agree.
I do not reward or support incompetence when the party in question is demanding my money for their product.
Why do people attribute decisions like that to the competence of the programmers? This is a business decision, nothing else. Most likely, some MBA looked over the numbers, saw a few hundreds or thousands of hours logged for tasks related to supporting Linux, and decided that Proton was good enough. Most likely, no programmer was even asked whether Linux support should be dropped.
And yes, even if you know what you are doing, every build going out to tens of thousands of active players needs to be tested, and that costs time and thus money, which is something every experienced cross platform developer should know.
Why do people attribute decisions like that to the competence of the programmers?
Because supporting multiple platforms, especially in gaming, isn’t magic or rocket science and almost always comes down to the setup of the toolchain.
This is a business decision
Very possible. But I go by their actual statement: “maintaining the native build across many distros was taking time away from developing new content”. My point is regarding the “maintaining […] across many distros” and not the “taking time away”. A good toolchain would make these differences extremely minimal.
hundreds or thousands of hours logged for tasks related to supporting Linux
Extremely unlikely. That would mean more than 10 developers working fulltime purely on Linux support since the release of the game. According to their team page on their website they have 7 developers in total.
every build going out to tens of thousands of active players needs to be tested
This is why experienced developers decouple the game from the platform specific stuff and test them separately.
The game is made in Unity so most of the platform specific stuff should already be production ready. Unity literally markets their engine as “Industry-leading multiplatform support” with the motto “Create once, ship anywhere”.
So my argument still stands. And as I said, it’s not a bad thing. The only thing I dislike is the indirect implication of Linux being a hassle when it would be nicer if they would take more responsibility for it.
comes down to the setup of the toolchain.
Unless you’re developing graphics-heavy application that uses platrofm-specific API for optimisation. Like a video game for example.
Are you talking from experience?
Not a game, but I was involved in making a graphics heavy app. Significant amount of times we had to grapple with question " that’s a nice feature you are making, will this work on all the platforms", and significant amount of times the answer was “obviously, unquestionably no”.
You should get some experience using modern game engines so you can see firsthand how easy it is to develop cross-platform games.
Supporting multiple platforms is a job for the engine developers.
And you should get some experience developing something more complicated than a simple unity project.
At my studio we maintain a native Linux version with a custom game engine, and it indeed takes a lot of time. I don’t consider Proton a viable option as we lost the ability to integrate with Linux-specific stuff such as Wayland APIs or better input, but I can definitely see the appeal of switching to Proton… if your team uses Windows. If you have some developers on Linux, you naturally get a Linux build (if using cross platform APIs ofc) and it’s actually faster to cross-compile a Windows build every once in a while (skip the slow ntfs I/O) and ship that. But it requires getting more of the team on Linux :)
But it requires getting more of the team on Linux :)
Get them a Steam Deck and target only Steam Linux Runtime 3.
Ah, yes… if only. I’ve upgraded internally SLR 1.0 -> SLR 3.0 but we can’t deploy it until a bug is fixed in the Steam client that causes, when we enable SLR 3, all Steam Decks to run the Linux build. Yes, Steam Decks run the Proton version, solely because the save file has different letter casing (yes I know it’s so annoying haha). We’ve spent quite some time on this and there’s no way to fix this without some folks losing their saves, and that is absolutely not an option. Soooo for now desktop Linux is stuck on runtime 1.0, and Steam Deck users are stuck on Proton. “fun” :/
we can’t deploy it until a bug is fixed in the Steam client that causes, when we enable SLR 3, all Steam Decks to run the Linux build. Yes, Steam Decks run the Proton version, solely because the save file has different letter casing
Sounds more like the bug is on your side caused by whoever had the genius idea to use different file names for Windows and Linux builds.
there’s no way to fix this without some folks losing their saves, and that is absolutely not an option.
For me the fix looks very easy: Use completely new file names. The old saves are only read and new saves get the new file names. Stay on the current SLR/Proton setup until the saves are migrated to the new naming format for the active user base. Make a dedicated “beta” branch for legacy saves in a year or so to not screw over inactive users. Make announcements and pinned forum posts.
On the one hand, it’s a shame in general, as Proton has truly been a pesky thorn on the foot for Linux gaming. There’s a world of difference between having native, first-class support, and basically running every game on an emulator that is on a lease.
