Ethan Lee, the developer responsible for the XNA reimplementation FNA along with around 50 game ports to Linux and macOS has announced the sunsetting of future macOS ports with a big update to a bunch of existing games.
Don't know who Ethan Lee is? You've probably played plenty of games either directly ported by them or running on tech created / maintained by them including: Streets of Rage 4, Superliminal, FEZ, Transistor, Rogue Legacy, Salt and Sanctuary, Owlboy and loads more.
Writing on their IcculusFinger profile page, Lee wrote a few paragraphs about a little history before diving into the current state of Apple products mentioning "I've put it off for as long as I could, but after looking at Apple's trajectory vs. my own infrastructure for Mac support, it's looking like 2021 is the year that I have to say goodbye to the Mac as a primary target.".
The good news for Linux fans? According to Lee, "New games, however, will primarily be Linux (and Windows, if applicable) only.". It also seems quite interesting that console releases have been opening up the gates for Linux ports according to Lee too. This is also backwards to what we often see with Linux support removed (hi Rocket League), so it feels a little odd.
As for existing ports? Lee went ahead and updated over 30 currently live with the latest tech for both macOS and Linux, so you should see the games ported all continue running smoothly in future.
Hopefully Ethan Lee will have many more Linux ports planned for 2021 and beyond.
I also found the bit about drivers interesting, with how users and developers can grab a RenderDoc capture and post it to the Linux Mesa driver issue tracker when encountering a driver problem and sometimes get a response and a fix in the same day - shows how awesome open source can be, compared with Apple "reporting bugs to Apple when you're not rich is like pulling teeth with boxing gloves".
On top of all that? The latest releases of FNA, FNA3D and FAudio are out now.
Quoting: EikeI just wondered yesterday, reading about Apple's new ARM processor: What will Steam be doing?
It will be interesting to watch what many companies will be doing... like Adobe for example...
Quoting: pbQuoting: EikeI just wondered yesterday, reading about Apple's new ARM processor: What will Steam be doing?
It will be interesting to watch what many companies will be doing... like Adobe for example...
As far as I understood, Adobe support is coming, albeit a bit late. Generally speaking Audio/Image and Video world will most likely continue supporting apple, unless they get trampled in the professional world. They have a lot of leverage there.
For Steam, honestly, I never understood why even bother to port games to MacOS. Mac is already not such a big market, but the number of Macs that can actually game is probably multiple times smaller than Linux gaming market. Macbooks are KO, mac mini is KO, so it leaves you with a bunch of iMacs and Mac Pros users I guess ? I mean, what's the point ? Of course, now that vulkan support (through moltenVK) exists, and since it is posix, it might still make some sense to support both Linux and MacOS, but MacOS as a primary port platform makes no real sense to me.
Btw, if anything, with Valve not giving any effort on Proton/Steam play on Mac, I think we kinda have our answer: they won't shut down the remaining lights right away, but I think they have already disengaged mostly. As long as there are people buying games there, they will probably keep it alive, but I would not expect anything more.
Last edited by 3zekiel on 4 January 2021 at 1:18 pm UTC
You either go all-in with Apple, or you really have to carefully consider if it is worth it? Even though Apple still has a cult following in the developer community, I think they are hugely overestimating just how far are people willing to follow them.
In the end, if Apple doesn't care about gaming, why should game devs care about Apple?
Quoting: Ethan Lee... reporting bugs to Apple when you're not rich is like pulling teeth with boxing gloves.
LOL! What a great simile!
I mean, I can understand flibit's position here, one person holding down the fort on their own is hard.
QuoteI'll even attempt to build Apple Silicon binaries and support that hardware as well if it seems like it's easy to do without first-hand testing (admittedly a bit lofty to attempt, but considering I'll probably have to support Aarch64 on Linux/Windows soon anyway it doesn't seem crazy).
This is great, though. Supporting more architectures is always a noble goal.
It will also help find some bugs in the code that are hidden by x86 idiosyncrasies, including reordering constraints and memory alignment. Though probably not immediately on current Apple Silicon machines, since they're still able to run x86 binaries through Rosetta 2. I fully expect Apple to remove that feature in the near future, though, this is what they did with the first Rosetta too back then.
Quoting: DrMcCoyThat's sad. While I personally don't use macOS and sometimes macOS vs Linux is pushed into some kind of rivalry, I'm one of these "radical portability" types. I want more portability, not less.
I mean, I can understand flibit's position here, one person holding down the fort on their own is hard.
QuoteI'll even attempt to build Apple Silicon binaries and support that hardware as well if it seems like it's easy to do without first-hand testing (admittedly a bit lofty to attempt, but considering I'll probably have to support Aarch64 on Linux/Windows soon anyway it doesn't seem crazy).
This is great, though. Supporting more architectures is always a noble goal.
It will also help find some bugs in the code that are hidden by x86 idiosyncrasies, including reordering constraints and memory alignment. Though probably not immediately on current Apple Silicon machines, since they're still able to run x86 binaries through Rosetta 2. I fully expect Apple to remove that feature in the near future, though, this is what they did with the first Rosetta too back then.
Having portable code is indeed quite cool, same for multi arch. But you also need to balance gain vs cost. Supporting windows + Linux + I think Switch for this case (I seem to remember he also works on Switch stuff) is already plenty enough.
For MacOS, I do think it should be done if it does not cost much, but I think the issue is overall Apple is a burden. The part about bug report only for riches is a good example. It's always the same in the end, gain vs cost. Being multi plat will help stability up to a certain point, but it's a diminishing return. You find lot of nasty bugs by adding a platform, much less with the 3rd one, and then less and less. If you have console + windows + Linux, MacOS will not add much more. ARM porting maybe, but likely not much either. Alignment and co should be managed by the compiler unless you go quite low level, at which point your code is likely arch specific.
See more from me