M4 Macs can't virtualise older macOS

Commenters lament that Rosetta will go away before users are ready.

In my opinion, Rosetta should be more heavily gated* to push everyone from Adobe to Steam to build for aarch64. Countless "Apple Silicon native" claiming tools require Rosetta under the hood for dependencies or even (bless their hearts) only their installer.

* Like right-click to open packages or install not-from app store, except Rosetta dialog should note that once it's installed any other old software not made for this system will run without warning. Turns out avoiding Rosetta is a great way to ensure not just apps but your CLI tools are all up to date... Alternatively, make Rosetta sandboxed with apps, and/or let it be turned off again more reasonably than the safe-mode surgery required now.

> In my opinion, Rosetta should be more heavily gated* to push everyone from Adobe to Steam to build for aarch64.

Honestly I see game developers simply abandoning MacOS (more than they already have) if they need to do more work on their back catalogues. Nobody at Adobe cares if CS3 doesn't run on the latest MacOS, it's no longer a revenue source for them. EA would care a lot more if all their games older than 4 years needed work to run on MacOS, there's a lot more long tail revenue in games (but a lot of that is over aggregated back catalogues where the revenue per title is pretty low even if they sum up to a significant amount) than there is in the more typical B2C cloud SaaS attached apps that the App Store model is designed to serve.

The scuttlebutt is that Steam is not going to get ported, and Valve has given up on Apple since the 32 bit drop in general and the Metal/Vulkan mess.

I have seen a huge downtick in games with Mac releases too… even for stuff where it seems like it would be possible with an extra platform export.

Apple seems to care about gaming for about 15 minutes every year, and one day they will figure out it can work on their stuff if they’re willing to accept that their platform will be nobody’s priority.

> the Metal/Vulkan mess

But what about MoltenVK?

> to push everyone from Adobe to Steam to build for aarch64

Not happening. The other commenters are right - many developers are content letting their apps get abandoned if the choice is between Universal binary or x86. A lot of software, particularly legacy software and older games, don't even have the opportunity to build for aarch64. The moment Apple put an expiration date on Rosetta they were confronting you with the inevitability that your software would one day die. There is no convincing some people - for christsake, four generations of Apple Silicon came and went and Steam is fine leaving their client x86 only. They know all their MacOS users are playing through Game Porting Toolkit's Windows version anyways.

From where you're standing, it must feel like an 18 carrot run of bad luck. Truth is, the game was rigged from the start.

What about Game Porting Toolkit? To be honest I feel like that's resulted in fewer native ports and more Mac players enjoying the x86 library that would never be ported natively.

Going from DirectX to Vulkan to Metal is simply too many translation layers to work well. You're almost always going to end up with an annoying bottleneck and poor hardware utilization. MoltenVK alone might be fine if Vulkan was more widely supported by shipping games.

I think it's more plausible that Valve decides to make a Proton for Mac using the D3D to Metal translation layer from Apple's Game Porting Toolkit—but that would be going against Apple's intended purpose for the toolkit.

The Linux Steam client is also still 32-bit, so it's more like 20 years with no 64-bit support.

I don't know, this is just stuff I heard, but I think some of it is Valve just being _annoyed_. Like Apple made this choice despite reaching a point where the highest end graphics card they ship is equivalent to a high-end... laptop graphics card! You shouldn't be in charge of a thing if you're not going to cover a wide spectrum.

At least with USB-C Apple "donated" their designs.

I don't see how. The performance of Game Porting Toolkit is in my experience very poor.

It's ok for older games but those would be very unlikely to receive a native port anyway, and for anything new GPT just sucks...

Why should all other companies and developers spend money to help $2T company make even more money when it was that company that broken all the existing software? Apple users will through their money at Apple every opportnity they get but won't spending any money on new version of 3rd party software that Apple has broken with their changes. I would abandon Apple if had any software released on their platform.

How is Mac no longer a revenue stream for Adobe?

Poster specifically mentioned CS3, which was a perpetual license. Adobe is not incentivized to keep a version of their software someone purchased once seventeen years ago working when they would much rather sell a monthly subscription.

The final straw for me was Apple deprecating modern OpenGL. There's no quick and easy solution to migrating even to Metal. So, they shafted us, and I'll never recommend another game developer bother with an Apple device ever again.

Use Windows for games.

But wasn’t macOS the only major gaming-relevant platform still using OpenGL anyway? If you’re already writing a Mac-specific backend, why not just write directly for Metal instead of OpenGL? Or, at the least, write for Vulkan and use a translation layer like MoltenVK?

> At least with USB-C Apple "donated" their designs.

What do you mean? I've always thought that it was designed jointly by the same standards committee as the previous iterations. The first USB-C device I know about and have used was the Nexus 6P. This was early enough in the standard's life that no one had any USB-C cables (or had any idea that it's a thing) so I had to carry my own one at all times in case I wanted to charge my phone. Apple started putting USB-C ports into MacBooks a year or so later, iirc.

Apple don't care to make games compatible, valve can't fix their design decisions why do you think theyare investing so hard in Linux, while Mac is already half the Linux userbase

> But wasn’t macOS the only major gaming-relevant platform still using OpenGL anyway?

At the time, there was a decent number of non-gaming cross-platform applications that relied on OpenGL for rendering. OpenGL wasn't perfect (especially compared to DirectX 9) but it was a good-enough solution for simpler apps and games that wanted the write-once-run-anywhere treatment.

> If you’re already writing a Mac-specific backend, why not just write directly for Metal instead of OpenGL?

Because a lot of people don't write Mac specific backends in the first place. Unless the app was designed to be Mac native from the start (a rarity in the professional world), there is very little impetus to rewrite everything to work with Metal and/or AArch64 targets. OpenGL suggested a future world where this would be unneccesary, and people liked it. With Metal as the only option now, a lot of people feel like Apple slammed the door on people that wanted to write cross-platform apps supporting Mac.

> Or, at the least, write for Vulkan and use a translation layer like MoltenVK?

MoltenVK is too slow for games (compared to DXVK it is an utter slouch) so most people don't even bother. There are a few apps that you can run in it, but for the most part it is a toy that rightfully isn't relied upon to deliver industry-standard experiences.

My understanding of the narrative was that Apple showed up with a lot of stuff regarding USB-C at the outset, even if they weren't the first person with it on their device.

This was just stuff I heard though. Vaguely, many years ago.