Hadders wrote:
I guess nvidia have done the easiest thing they can, which is provide a base and let the fanatics do what they will.
Of course releasing specs for those developing upstream projects would trickle down to all distribution variations by those that compile the programs that are based on the upstream kernel and GUI developers efforts.
There is nothing fanatical about not purchasing products that do not taylor to assuring or at least giving effort that their product will not fry or cause instability on any user's system.
I have the card because it was in the junkpile for parts I was given by another person. The funny thing is that I got rid of a prior nvidia card because of the lack of 3D OSS drivers available.
Anyway, nouveau (OSS 3D driver under development), the 2D nv driver included with the GUI, the scripted binary driver with a hook or not use any of the products that they sell with no intention to assure users that they work without disrupting an otherwise stable system.
Fanatical, rational or principled doesn't really matter. The driver is closed source from the supplier. You can do whatever you choose. The question was mainly why one form of the driver is better than the other. The posting that referenced Mike's feedback on the difference was pretty much what I tried to convey. I think the content was close, but the wording totally different.
Some kernel developers, one not currently working for Red Hat have commented regarding the effects of NoSourceDrivers and bug reports. Without the source, how can one determine reliably what is happening when the binary driver does bad things with otherwise stable systems.
Jim -- Your talents will be recognized and suitably rewarded.
