Why not make it long term?

From what i know ogl was the best graphics api for 10 years with modification imposed by the new capabilities of the graphics chips. Could’n it be possible to create an api that will be independent of the future modifications of the graphics chips?
I mean, could ogl2.0 be designet in such a way that it will support psX.0 and vsX.0 before they apear, and the programmer to design only diferent versions of shaders (eg 1.4 for 1.4 capable cards, 2.0 for 2.0 cards , and so on), no new extensions, no nothing, a clean and straight forward api? Just a thought…

There’s more to an API than just shaders.

Yes, they can design a shading language that would cover expected functionality. However, that’s not good enough, because you never know when the next unexpected thing might come out. OpenGL had no facility for shaders, and it cost them.

Also, building such a shading language has certain problems. The principle problem is that it is so forward-looking that it can’t run on modern hardware, and thus becomes a useless API. One of the problems with early OpenGL was the difficulty of trying to figure out what functionality would bounce you into software and trying to code around that. You don’t want to put that into a shading language.

As for fixing the rest of the API, it doesn’t need fixing so much as cleaning up. Removing functionality that has been superceeded by others, and so forth.

Fundamentally, you can’t make an API that is nice to use that can service all future needs, even if you don’t know what they are.