Irrlicht Engine

You can write both animations/animators without using full-blown physics implementations. As you won't want to spend most of your time on calculating millions of particles flowing in the air/water, it's an approximation anyway. And it'd be your task/thesis to develop a nice effect which looks as much as the desired physical effect without needing all those complex calculations. And of course you can have interactions or alterations such as changing wind. Just add some more parameters to your animator, and let the user change it during run-time.

http://www.dgp.toronto.edu/people/stam/ ... /GDC03.pdf
http://paper.ijcsns.org/07_book/200805/20080513.pdf
I saw a surface fluid simulation on GPU gems too (not the 3D fluid simulation from GPU gems 3 which is more complicated).
There are many things on the web about this kind of simulation.

In fact I don't talk about 3D fluid, where you use a lot of particles and a lot of physic, but fluid surface.

I still don't know with whom you're discussing, stefbuet, but I don't talk about such simulations either. As I might have made clear in my previous post.

hybrid wrote:I still don't know with whom you're discussing, stefbuet, but I don't talk about such simulations either. As I might have made clear in my previous post.

Thank you hybrid

I understood what you said, but not completely, and I know that there will be no physics for Irrlicht. I will try to see what can I do with animators.

I was just explained what I was talking about, that is too say fluid surface and not 3D fluid flow. The first one can be achieved without any full physic engine. As I was saying it, I will probably do that next year : a scene node like the waterSceneNode, which water surface is dynamic, I mean you can produce some ripples with the mouse cursor... or with another mesh node. I've only looked for somes documents about it, but still need 1 more year to have all physic knowledges required to do it

stefbuet wrote:I was just explained what I was talking about, that is too say fluid surface and not 3D fluid flow. The first one can be achieved without any full physic engine. As I was saying it, I will probably do that next year : a scene node like the waterSceneNode, which water surface is dynamic, I mean you can produce some ripples with the mouse cursor... or with another mesh node. I've only looked for somes documents about it, but still need 1 more year to have all physic knowledges required to do it

So you think it will require one year physics knowledge to have the ability to implement it? How much time do you think it will require for implementing it along learning knowledge to do it?

I personnally need 1 more year before before being able to play with fluid physics. Are you full time on this project ? If you have never done some physic stuff before with differential equations for exemple, maybe it could become difficult. You need a physic knowledge background to anderstand that. I can't tell how much time you need, I never done it, and haven't started fluid lessons at school yet.

I think fluid physics suggests that we're straying a bit from Irrlicht's role as a graphics enigne, not a game engine. Wouldn't something like postprocessing, Cg support, DX10/11 support, or the vertex reworking be more beneficial to the engine's longevity and usefulness as a whole?

Personally I'd love to see postprocessing or official Cg support, since Cg would play very nicely into the multi-platform and multi-renderer nature of Irrlicht.

How about building a grass system like that in GPU Gems 1? It does not use physics and irrlicht will remain graphics engine. I think I can implement it by vertex shaders for both GLSL & HLSL.