Hello, On Mon, Aug 13, 2007 at 06:18:10PM +1000, Hugh Fisher wrote: [...] > Abandoning the fixed function pipeline is just recognition > that all the action is happening with programmable GPUs > now. Since not everyone is writing shaders though, OGL 3 > will initialise the graphics hardware with shaders that > emulate the original OpenGL pipeline. People who write > OpenGL 1.X code, eg for OGC 1, won't notice any difference.
I admit I haven't read much about it yet, but isn't a large part going to be fixed-feature still? E.g. I'm really certain that triangles will stay the main "object" that is handled... > Same for glBegin/glEnd blocks being removed. In the early > 90s it made sense to transfer vertex attributes (xyz coords, > tex coords, etc) to the graphics hardware immediately they > were available, ie each glVertex call. Now we have an order > of magnitude more RAM, it makes sense to buffer an entire > glBegin/glEnd block into a vertex array or vertex buffer > object and send it in one block. With most OpenGL systems > already doing this internally, removing glBegin/glEnd is > just hitting us programmers with the clue stick :-) But > again, people who write for OGC 1 won't be affected. I think one of the arguments was that even if the hardware does not support vertex buffers or similar it is still faster to use a highly optimized function in the driver that uploads the array one by one than using thousands of function calls. The only exception is of course when you only have very few coordinates/triangles, but then speed really does not matter anyway. Greetings, Reimar Döffinger _______________________________________________ Open-graphics mailing list [email protected] http://lists.duskglow.com/mailman/listinfo/open-graphics List service provided by Duskglow Consulting, LLC (www.duskglow.com)
