I'll just leave this here... http://www.anandtech.com/show/8363/khronos-announces-next-generation-opengl-initiative

In short, Khronos are creating a new next generation graphics library which will allow explicit control over CPU and GPU workloads, similar to that in Mantle, D3D12, Metal, etc. Some of the key features are:

• Ground-up redesign of the API
• Intermediate language for shaders
• Explicit CPU/GPU control
• Unified graphics library for desktop and mobile
• Support and design contributions from some major game companies - Valve, Epic, Blizzard, EA etc

I've already said this in a few places but I'll repeat it here...

I find this amusing more than anything because for around the last year now OpenGL involved IHV representative have been yelling that 'Approaching Zero Driver Overhead' is enough and that 'we don't need another API' because OpenGL was low level enough and with the AZDO extensions (which only work on a couple of bits of hardware right now, but lets skip over that) you could avoid all that nasty overhead.

This line was being trotted out right up until recently, where GL was fine and we didn't need something new and everything was awesome! (to quote a certain Lego film I've not seen but everyone keeps quoting at me).

Mantle wasn't needed.
OpenGL was good enough on it's single thread!

And then today we get this.
A new API.
And an announcement which says work is already underway.

So all the 'OpenGL was fine' was clearly rubbish; they knew it, we knew it, so they should have just admitted it instead of constantly trotting out the 'AZDO is the solution' line which they beat us over the head with.

While the news is certainly interesting I just find the whole attitude for the last year disappointing; they should have been saying 'here is part of the solution' and even if it wasn't up and running yet at least indicating they were considering things. If nothing else it might have stopped people yelling at them if they could admit their API wasn't up to the task when it comes to modern multicore hardware and the workloads game devs want to throw at it.

Also, having lived through the OpenGL2.0 and OpenGL Longs Peak/3.0 debacles I'm very much in a 'believe it when I see it' frame of mind; it stands a better chance given the other changes in the API landscape, and some good people are involved... but in their previous guise they have screwed up before so... we'll see how it pans out.

+1 on phanotom

Not sure whether to laugh now or afterwards.

*excitment increases*.

seriously, why does their seem to be a bit of negativity in the air? for me personally this is what i've wanted to see for quite awhile, get rid of the old cruft, and build something that is more in line with modern gpu and cpu pipelines.

The idea is fine enough, but they have screwed up twice before and while this time it might stick they just spent the last year saying "the API is good enough, we don't need another" which is what they are now proposing.

At this point its less about excitement for me and more 'hey, welcome to the party at last, I see you've finally worked out where it is...' given that we've had 3 API announcements (and two launches) in the last year and they have only just got around to saying they are looking into it.

Mantle and D3D12 are looking to be largely the same, the same people are involved in this (from the ISV side) so it's unlikely to suck API wise and will, more than likely, look something like Mantle.

We also got GL4.5 with Direct State Access! Finally!

OK, it only took the ARB 6 years or so to standardize DSA, despite the fact that virtually everybody was a-hollerin' for it. It's also notable that they didn't adopt the parts of DSA that relate to deprecated functionality (which was expected) despite noting that the new GL_ARB_direct_state_access only requires GL2.0 (which was unexpected).

What's quite neat about the latter is that it's now possible that we'll see faster uptake of this extension by Intel and AMD drivers who don't need to wait for a full shipping GL4.5 driver.

The idea is fine enough, but they have screwed up twice before and while this time it might stick they just spent the last year saying "the API is good enough, we don't need another" which is what they are now proposing.

Communication is most definitively not Khronos' strength, you should already know that from when they brought out OpenGL 3.0

-- same story really. But in retrospective, it was a good thing. Maybe not as good as it could have been, but still. I see this as good news.

Communication skills 0/10 yeah... but still good news.

I am making a game in OpenGL "for fun" right now. I have absolutely no need for this, as I'm already running at about 2000 fps when vsync is off.

I hope they don´t turn it into some AAA-game only api, but keep some higher level stuff for us who just want to draw stuff to the screen.

OpenGL as it is now won't suddenly disappear. It'll still be there for people to use.