That's a good point, but when/if the day comes that DX12/13 offers benefits like mantle and works with all platforms, it wont be a great situation for mantle. Its cool that AMD support both mantle and DX12, but I don't really see it as the best business move ever.
directx will never be available to anything other than MS os's. its Microsoft's only real leverage to make you buy the OS.
If it wasnt for direct x being MS only Most people here would probably be running linux and im sure they would have made distros that even a newbie could use if such an exodus of MS users ever wanted to use linux, because it became a viable gaming platform (look at steam os for instance)
MS would be in a really bad position if that happened..
they don't have enough market share any where else "IMO" to be able to wave good bye to their bread and butter.
mantel or an improved open gl is our best bet for a more level playing field when it comes to OS choices.
open Gl is just a mess though. developers do not want to program for open gl any more due to all the years of added code and stupid programing rules, open gl would need a complete over haul and clean up for it to be viable.. which is a shame because really it has always had low level hardware access which is what mantle and dx12 are claiming is so great about their new api's..
having said that dx 12 will also come with more work. previous direct x api's would do a lot of work for you releasing resources and so on automatically. this however slowed it down a bit and made it a bit less efficient.. with direct x 12 they stripped all that out so programers will have to make sure that they have it using and releasing things as and when needed which will be extra work. and possibly make it more like open gl. I dont know much about mantels usage of resources though or how you would go about allocating them or if it releases them its self /you have to code it specifically to release them..
The main problem with mantel IMO is amd wouldn't let any one have a look at it when they asked for access to it to start it. even though they claimed it was open source. nvidia and intel asked and were denied. Which is not a great move.. i have a feeling if amd had just let them have a look at it when they asked then it would have had more support, but i guess "and i don't know" that amd wanted to do more work on it so it worked better when they let them look at it "or even make sure it worked the best it could with their own hardware". but if they had shared it when asked then every one could have had a look at it and made it work better for everything.
mantel could well end up a bit like 3dnow and end up being amd specific but cross platform. this could at least let Some games work on linux/macs and so on. but then it would be a small selection. i dont see game devs coding for 2 or more api's as a default. the only way it would have a chance is if the xbox and play station used mantel as its api. but we already know the xbox will be dx 12. which just further limits mantels scope.
)