> The chances of Apple pulling off what Khronos couldn't are slim - in my opinion.
Apple has a large and loyal customer base, and it's typically people who are willing to spend money on apps and games for their devices. Apple doesn't have to try and satisfy a bunch of legacy CAD vendors who don't care about consumer 3D graphics at all. And when Apple announces Metal, they don't mean they have a spec, they have an implementation that is going to have a significant installed base in under a year. They have a lot of advantages Khronos doesn't.
OS X has a lot going for it, but when it comes to the desktop gaming market it's just not the top dog in revenue generation that iOS is for mobile apps. I'd assume (not an expert) that Vulkan's support from Intel, nVidia and AMD on PC could mean more to professional game developers and publishers than Metal's OS X support.
I don't think there's going to be a major game developer that supports OpenGL/Vulkan on Windows over DirectX, outside of perhaps Bethesda/iD. If Vulkan isn't present on the second-most used desktop OS, and it's not best-of-breed on the operating system used on over 90% of desktops, I don't see where it makes major inroads on the desktop.
Valve is certainly interested in Vulkan, partly as a means to target Chinese Dota players running Windows XP. Even a second-place position on PC may amount to more than star status on OS X. (Though OTOH the relatively easy port could draw people looking for some extra income for their iOS game.)
(In any case, is it yet certain that Apple won't support Vulkan on OS X? It wouldn't have been ready for El Capitan's launch this Autumn regardless. Maybe Apple will try to deprecate it (in the non-technical sense) as a compatibility option but support it anyway?)
Valve is an oversight, yeah -- I did mention them in another comment here. I don't really think of them as a game developer anymore, but I'm probably being unfair.
If Apple supports Vulkan but with worse drivers than Metal, I'm not sure that's any better. OSX has roughly four times the penetration of Linux in the latest Steam hardware and software survey. I'm not sure it's worth it for most developers to deliberately target the second-best choice on OS X in order to pick up the handful of Linux gamers who can't/won't dual-boot Windows.
I was talking more about Vulkan's second-place position on Windows than about Linux (Vulkan is likely to have some advantages on Windows, like being less tied to the Windows roadmap than DirectX, or indeed OpenGL), and if Apple provided bad Vulkan support I think they'd just be letting themselves in for bad Vulkan-based ports. But overall I don't think we're really disagreeing very much here.
Did you mean Vulkan where you said "Mantle's second-place position on Windows?" As far as I can tell, Mantle is pretty much DOA -- Nvidia and Intel aren't going to support it on any platform, and Nvidia has 52% of the PC market according to the latest Steam user survey[1], and Nvida/Intel combine for roughly 80%. Even AMD is deprecating Mantle in favor of DX 12/"GLnext"[2].
Yes, of course I meant Vulkan, sorry. IIRC though Vulkan is effectively Khronos-standardised Mantle, so it's not all that fair to say that Mantle has died exactly.
> Valve is certainly interested in Vulkan, partly as a means to target Chinese Dota players running Windows XP.
That seems like a far fetched theory to me. AMD and NVIDIA haven't released drivers for XP for ages. Vulkan support in Source Engine is nowhere near done, if it has even started. In the meantime Chinese players will move on...
Apple has a large and loyal customer base, and it's typically people who are willing to spend money on apps and games for their devices. Apple doesn't have to try and satisfy a bunch of legacy CAD vendors who don't care about consumer 3D graphics at all. And when Apple announces Metal, they don't mean they have a spec, they have an implementation that is going to have a significant installed base in under a year. They have a lot of advantages Khronos doesn't.