Apple will switch to ARM processors: what does it mean for plugin developers?

[BONES]

Windows has tried to do this with a spectacular lack of success. The Microsoft Store is a complete joke, utterly impossible to take seriously. More to the point, Microsoft knows their users won't put up with that kind of shit, just as Apple knows its users will put up with whatever shit Apple decides to throw at them.

[Michael L]

Apple knows its users will put up with whatever shit Apple decides to throw at them.

Not quite. Apple limits choice to make everyone follow the current plan. Common example: a recent security update introduced a recognised firmware bug but Mojave was 'end of life' in October so the bugfix was only for Catalina. I spend weeks with Exec Relations trying to change that policy to match reality, but its too top-down. Steve Jobs once said, "You don't know what you want until we show it to you." That's Apple in one sentence.

[IncarnateX]

when macOS moves to ARM, will vendors still be able to sell their own software or will Apple want to force everyone to use the App Store going forward? i.e. Will Apple's insatiable greed mean they are happy to lose products from their platform in order to get 30% of whatever's left? To me that seems like it might be one of their goals.

It would not come as a surprise to me either. They may be going for a “natural selection” here.

Further, if Apple is going for a unified app store between mac and iOS, it would be more in their spoil-the-user-at-the-cost-of-devs-spirit to try to lower the prices of the mac apps instead of raising the iOS apps to the prices of the mac store. They are somewhat 1:5-10 for some apps, right? However, then they would have to rip off the devs instead and question is how much more they can squeeze from these without draining lives?

[mystran]

It's not just a matter of hardware, because macOS itself is rather hostile environment for games. Even something as basic as raw mouse input is literally impossible without kernel extensions, which means that for any sort of action gaming involving the mouse, the same machine will provide a better experience just by dual-booting into Windows.

It's a bit of a pain, but you just need to subclass a view to get your mouse and keyboard input. However, now that OpenGL is deprecated, I'm not sure how one would do this anymore since the OpenGL view will no longer be an option.

I'm talking about raw mouse input.

Basically on any other operating system, there is an API to get the uncooked data from the mouse directly. The operating system will not adjust the speed (which reduces precision, because the cooked value is rounded to pixels), will not perform any acceleration (which is a total no-go especially for FPS; Apple also has particularly crazy acceleration in general) and will not limit the mouse movement into the window bounds (which is a problem for 3D aiming). It literally just gives you the raw delta-ticks that the mouse itself reports.

You can certainly limit the mouse into a window and then warp it in the middle on every update, but this doesn't solve any of the above problems: the input is still cooked, OS acceleration is still applied and sufficiently fast movement will get clipped into the window (or screen) bounds. In fact these problems are precisely why every other operating system added raw mouse input decades ago.

ps. Note that such an API has nothing to do with the "mouse" as we think of it when it comes to regular GUI programming, but rather it's about raw devices. For example on Windows, if you have multiple mice connected, the game can even tell which mouse is sending which input, so you could build a mouse-driven local multiple-player game with multiple cursors if you wanted to. You can also use raw input for keyboards too, but that is usually not quite as critical as those don't get cooked so hard.

[arne]

You can use the IOKit.framework to get raw data from any HID device connected to the system. This is as raw as possible.

[mystran]

You can use the IOKit.framework to get raw data from any HID device connected to the system. This is as raw as possible.

Right and then you need to basically write a driver for every device around there, because plenty of them need specific workarounds (see Linux source).

[arne]

I never actually tried it, did you?

[mystran]

I never actually tried it, did you?

Actually I kinda did. I spent 2-3 weeks on it and eventually had to give up (and I mean the whole idea of shipping a game on macOS, even though I mostly write code on an MBP these days; the stuff you have to go through is just not worth it)... and the reason I'm even writing game code is mostly just because I kinda needed a break from plugin API politics.

[syntonica]

I'm talking about raw mouse input.

Basically on any other operating system, there is an API to get the uncooked data from the mouse directly. The operating system will not adjust the speed (which reduces precision, because the cooked value is rounded to pixels), will not perform any acceleration (which is a total no-go especially for FPS; Apple also has particularly crazy acceleration in general) and will not limit the mouse movement into the window bounds (which is a problem for 3D aiming). It literally just gives you the raw delta-ticks that the mouse itself reports.

You can certainly limit the mouse into a window and then warp it in the middle on every update, but this doesn't solve any of the above problems: the input is still cooked, OS acceleration is still applied and sufficiently fast movement will get clipped into the window (or screen) bounds. In fact these problems are precisely why every other operating system added raw mouse input decades ago.

ps. Note that such an API has nothing to do with the "mouse" as we think of it when it comes to regular GUI programming, but rather it's about raw devices. For example on Windows, if you have multiple mice connected, the game can even tell which mouse is sending which input, so you could build a mouse-driven local multiple-player game with multiple cursors if you wanted to. You can also use raw input for keyboards too, but that is usually not quite as critical as those don't get cooked so hard.

There's the deltaX/Y/Z values you can poll. Sorry I didn't realize that these were pre-scaled by the OS, although there's no reason why a game couldn't rescale them back. I've never had to concern myself with anything beyond that.

I also thought you could get raw mouse/keyboard data from one of the Cores. I didn't know you had to go all the way back to the HID Manager. That really sucks.

[machinesworking]

Windows has tried to do this with a spectacular lack of success. The Microsoft Store is a complete joke, utterly impossible to take seriously. More to the point, Microsoft knows their users won't put up with that kind of shit, just as Apple knows its users will put up with whatever shit Apple decides to throw at them.

Microsoft has always sucked at branding, execution and anything to do with creating hype. They never had to, exclusive corporate deals will do that to your company. Microsoft are great at making products, they suck at selling them, it's no wonder their store isn't a success. It has nothing to do with any of your idiotic preconceived notions of what the end users are like.

[BONES]

It absolutely does. Compare the reaction of Windows users to the problems introduced in Vista to Apple users' reaction to the cluster-f**k that was Catalina. Windows users simply refused to update Windows until Microsoft fixed things for them where macOS users were happy to make excuses for Apple. On the Windows side we saw it again with Win8 and in both cases Microsoft was forced by it's user backlash to back-pedal and rethink what it was doing. Can you imagine Apple ever having to deal with a user revolt?

I agree though that Microsoft has shown a complete inability to sell anything to consumers. Look no further than Zune to see just how pathetic they are at it. Zune had a streaming service in 2009, at a time when Spotify was just an idea in some Swede's head. They had a massive headstart on the entire market but still managed to fail utterly, abandoning the venture completely at a time when streaming was just on the cusp of dominating the way we consume music. A company like that really doesn't deserve to succeed.

Apple knows its users will put up with whatever shit Apple decides to throw at them.

Steve Jobs once said, "You don't know what you want until we show it to you." That's Apple in one sentence.

That's also exactly what I said.

[Michael L]

That's also exactly what I said.

Mac users complain loudly as Windows users, but Apple is less responsive cuz its priority is the new shiny.

[adammonroe]

What it means for developers:

*Many weeks (and possibly months) of recompiling code and reuploading plugins for yet another unnecessary format/platform change.
*Many months of emails from users knowingly or unknowingly using the Rosetta emulator and asking why your software suddenly got 50% slower.
*Maintaining both Intel and ARM VST/VST3/AU/AAX plugin formats (4 extra versions, on top of maintaining 32-bit versions which I currently do for pre-Catalina legacy support and for people using older software like the 32-bit only version of Ableton).
*Many days of cursing Apple and wondering why the hell people keep buying Apple products when even users go through the same "uh, make sure to wait at least 6 months to upgrade to the latest OS because your software probably isn't supported yet" cycle as developers.
*A ton of wasted time on ports that could have been better spent on developing new plugins or improving existing ones. Honestly, I sometimes wonder how many hundreds of thousands, possibly millions, of developer hours Apple has wasted with their (mostly) pointless updates and changes throughout the history of software development.
*Another $1200 down the tubes on buying yet another Mac Mini or whatever for development and testing.

That is the other issue that needs to be clarified - when macOS moves to ARM, will vendors still be able to sell their own software or will Apple want to force everyone to use the App Store going forward? i.e. Will Apple's insatiable greed mean they are happy to lose products from their platform in order to get 30% of whatever's left? To me that seems like it might be one of their goals.

This is the line in the sand for me. If Apple ever mandates selling through the Apple Store I'm done. "Other people will come in and fill the void, you're leaving money on the table, blah blah blah" - yeah I don't care. Sometimes a man has to take a principled stand. If other people want to lick Apple's boots at that point fine. Code signing/notarization and being forced to pay for a developer fee is bad enough, no way they are ever getting 30% of my sale for doing absolutely nothing. It's already bad enough that they steal time - and thus money - out of my pocket every year with their constant "improvements." From a consumer perspective, I don't see how that doesn't artificially inflate the price of plugins to compensate for a 30% drop in developer revenue. Creating a closed platform would absolutely be the last straw for me.

[machinesworking]

It absolutely does. Compare the reaction of Windows users to the problems introduced in Vista to Apple users' reaction to the cluster-f**k that was Catalina. Windows users simply refused to update Windows until Microsoft fixed things for them where macOS users were happy to make excuses for Apple. On the Windows side we saw it again with Win8 and in both cases Microsoft was forced by it's user backlash to back-pedal and rethink what it was doing. Can you imagine Apple ever having to deal with a user revolt?

I'm not on Catalina for either of my computers. The cluster f**k, that to me, and I would guess a lot of other people could tell, was all about getting ready for the impending Arm chips. I don't think there was any mystery there at all.
The butterfly keyboard debacle is an example of hardware issues that mac users RAILED against, constantly. There are dozens of examples. The way I look at the upcoming Arm maneuver is good for them, it's a great business move, no doubt they will be successful at it. That said, I might not join them for it, audio is a niche market, they obviously with the exposure Logic gets like being part of it, but if developers are not able to port without a nightmare etc. then I'll just get a ZenBook and call it a day. More likely I'll eventually do what I always thought would be the optimum for my setup, a Windows desktop running VEP and and an Arm MacBook when they get as powerful as the latest 8 core laptops out now. The things likely will be dead silent.

I agree though that Microsoft has shown a complete inability to sell anything to consumers. Look no further than Zune to see just how pathetic they are at it. Zune had a streaming service in 2009, at a time when Spotify was just an idea in some Swede's head. They had a massive headstart on the entire market but still managed to fail utterly, abandoning the venture completely at a time when streaming was just on the cusp of dominating the way we consume music. A company like that really doesn't deserve to succeed.

Like I said, the fact they cornered the market early on has made them completely lazy about branding, they're beyond terrible at it. Their corporate culture is generic office culture, this is why you had the search puppy, and clippy.

You can ad Arm in there too, Microsoft currently sell a surface running on an Arm chip. We aren't spending dozens of pages talking about it because they didn't bother developing their own chip from what I know, and aren't aggressively getting developers to port to it. They're just sort of sitting on it with their thumb up their ass.

They have these fantastic developers, they get some great tech, then they just sort of tack it all together and call it good enough. is it still a thing on Windows that you have to specify an audio driver to a hardware USB port? that was just mind bogglingly silly to me.

[SampleScience]

For me, it's obvious that Apple will eventually force devs to sell in the App Store. With the current Catalina fiasco, the notarization problems, charging 99$/year to be able to sell software on macOS, and the lack of reliable support from Apple, I'll be forced to sell Windows only plugins in the next 2 years.

This is not fun but considering the scale of my business, I don't have much choice. This year alone I've lost 3 months of work just to make my plugin work on Catalina. That's money down the drain.

At this point, I'm not even sure it makes business sense since the plugins will only be compatible with macOS for 2 years and then I'll have to start all over again. Maybe I'm better off offering Windows plugins... or not:

Because going Windows only will create two problems:

1. the incessant harassing of macOS users asking for mac versions of my plugins.
2. much of press will ignore me because most music journalists use mac computers (from my experience)

The second problem is crucial, without the support of the press I'll never make anywhere near the amount of money I make now with Windows/macOS plugins.

For all these reasons, I'm thinking about pulling the plug and just do something else entirely. Maybe I'll just do sample libraries and that will be it. I don't know yet.

I was thinking about doing Kontakt instruments but the market is nowhere near as strong as plugins. Plus, it will probably be affected by Apple's decision to move to ARM too since NI will have to make Kontakt compatible with it. How long it will take, nobody knows. In the meantime, it's a shrinking market that might expand once the switch is made for Kontakt and everything runs smoothly on the new ARM computers.