Irrlicht 2.0 - What's in store for the future!

[Granyte]

DX11 is almost completely fixed the only thing remaining is the text scenenode is depth writing for no good reason.

But as i told befor the patch is already massive because the whole material system was poorly done and will probably only get worst when i fix the textscenenode

[lumirion]

In potential vulkan support and/or dx11 what about using a render target that uses per-pixel linked lists of fragments to provide order independent transparency by default. Any GPUs with over 100mb ram should handle it just fine.

[hendu]

Radeon R200 series, introduced in 2001, had 128mb VRAM cards.

[devsh]

In potential vulkan support and/or dx11 what about using a render target that uses per-pixel linked lists of fragments to provide order independent transparency by default. Any GPUs with over 100mb ram should handle it just fine.

Thats too high-level, I think its something the developer should decide or at least have very fine grained access to

Remember, memory is not primary concern, performance is.

[lumirion]

what about replacing type enumerations with some sort of type registration? That way deriving new extensions from the defaults would not require modifications to the Irrlicht source and the engine could dynamically accept new types. The cost would be slight over head of storing the type along with an int identifier.

Code: Select all

 
RegisterableTypeEnum
{
     string type;
     int ID;
     operator int*() {return ID;}
}
 

[kornwaretm]

how about scalable timer. lets say device->setGlobalAnimatorTimeScale(0.5); and all animation speed scaled down to half speed. device->setGlobalAnimatorTimeScale(0.0); and all animation paused, particle emiter stop emitting etc.

[CuteAlien]

@kornwaretm: ITimer::setSpeed should do that.

[kornwaretm]

ow it is already exist this whole time