below is the excerpt.

these are my reasons ;)
a.) i have no idea how carmack is doing this but iirc he has 3 or 4 (john scott told me it was at 4, but i e-mailed john C. and he only mentioned 3; that was more recent) code passes going on.
i have 3 sets of graphics functions for each card (ati, geforce, regular) but this is just getting crazy. i suppose c++ would have made this easier ;/

b.) i was somewhat fine with the above stuff, but then mark decided to plaster 72 point bold font down the center of every single page in the new NV_vertex_program spec!
explain to me how this is going to "promote an openness" ?
if i where at ati i would be afraid to even download the damn thing!
i know kilgard made a killing of the stock, as did i but still there is a limit.
money is good but there is the "right thing to do"
 

c.) extension specs are getting longer and longer. there are so many architectural  changes we are having to go through to support every "cool" gl extension.
 

d.) the whole extensions system started out as a way to "promote innovation and new ideas", But, apparently companies seem more bent on sueing each other then furthering the api...

e.) vertex shaders are damn fast! ;)
the spec is not to large and the instruction reference is simple and easy to understand.
unlike the NV_vertex_program
heh, mark decided to name it NV_vertex_program because shaders denote a type of fragment operation type.
heh.
i' mean c'mon, we are computer programmers we are known to name things wrong :)

f.) before vendors didn't have everything in hardware, there roadmap was to get the full specification in hardware.
now that they have done this, they are making there "own roadmaps".
in *general* this is not a "good thing" for us.

now we look at microsoft, they have "been" playing catchup, but they *do* have a "future" roadmap, and they talk with us all the time.

i converse regularly with mike and tony about "what to do next".

>accidentally posted by Pierre Terdiman
hehe, ;)
pierre also brought me to the darkside as well ;)
we had a discussion about d3d and GL.