Game development, Nebula Device, personal mumblings...
hehe :]Yep quite controversial!My opinion is: either extreme is bad. Absolute flexibility is crap, of course. Absolutely fixed and waiting months for shader features is crap as well. So I mean: plugged together Shaders have to be revised by a good programmer!Isn't thats a nice "beetween the chairs" viewpoint? As I am between graphincs and technical stuff.
good read. and i'm glad you've started posting so regularly.
Yes, I think the best approach is to have a relatively small number of shader "layers" (like ReflectionLayer, LightmapLayer, AnimatedUvLayer, etc...) which are provided by the programmers and can be "stacked" by graphics artists. Each layer has a "cost" assigned and as soon as a "layer stack" becomes too expensive, an alarm goes off. It's a bit like in the old multitexture-days really... It's especially important that components like lighting/shadowing remain centralized and "invisible" to the graphics artists.
Here is something else you might be interested in. Direct3D 11!http://www.bit-tech.net/bits/2008/09/17/directx-11-a-look-at-what-s-coming/
Post a comment