There's no point in upgrading Irrlicht to newer versions of OpenGL or Direct3D without fixing the main underlying issues that are preventing us from fully utilizing OpenGL 2.0 and Direct3D9.
If we are only using 50% of the features available to us in Direct3D9 then what is the point of upgrading to Direct3D10?
D3D10/11 support with deferred rendering etc. would be great.
What does D3D10/11 have to do with deferred rendering? You can do deferred rendering without modifying Irrlicht now if you have a bit of shader skill.
I think that if we update to Direct 3D 10/11 everyone will realise that everything still looks exactly the same from the outside and you still need to write shaders and graphics routines to get a nice look.
You have to be more precise with what you want, is it a marketing term or a specific feature that enhances the graphics? Most of the important changes made to D3D10 were things like code design and multithreading. Things that make it easy for the low level developer, but if you're an Irrlicht user than it won't make a difference, all the drivers have the same interface!
Things like geometry shaders and the SM 4.0 model are not really used, geometry shaders flopped and SM 3.0 already has a gigantic instruction count and if you need the amount that SM 4.0 is offering than it probably won't run that well (Or take ages to compile).
So in the end even if D3D10 was implemented it won't make a single difference to the end user unless we specifically add things that would make those features easy to use.
I have to ask why. What benefit does it give to the end user given the current level of support for already implemented APIs and exposed features?
]
