To me, the nouveau driver is old news. What WOULD be big news is information on nouveau's Gallium drivers that add 3D support to nouveau.
Radeon driver was already making very good progress and documentation has certainly helped but as Nouveau driver has proved, with enough focussed effort, we can jump through the hurdles necessary. Pidgin, Empathy et all support a crap load of protocols, almost all of them entirely reverse engineered as well.
In short, reward good behaviour with your money.
With regards to the ATI sub-thread, I'd like to note that nearly all of the current work put in on the Mesa and Gallium ATI drivers has been either commissioned through PI/TG prior to AMD's acquisition of ATI, or written by volunteers using AMD's docs.
Anyway, to answer the big question, nVidia will continue to maintain both nv and nvidia, and continue to ship nv in the Xorg katamari to the best of their ability.
Why, you ask? Simple. nVidia doesn't maintain nv for us; they maintain nv so that their customers (the people who buy their consumer-market cards) have an out-of-box solution that nVidia themselves can guarantee. Since they haven't written a single line of nouveau (save for the lines originally from nv), they can't actually make that promise without keeping nv.
For the same reason, nvidia will continue to exist, except that in this context, the customers are people buying professional hardware. Y'know, Quadros and Teslas.
Always annotative, never authoritative,
Ahh wait a second no he wouldn't, kinda because hes Gandhi...
I own an nVidia GeForce 8400M GS. ..I don't do as much gaming as I thought I would.
If nouveau can run Compiz with basic effects responsively and not crash, and has KMS.. I don't even care about framerates.
I bought an X-25M to get blazing fast (~4s total) cold boot times. It's looking like actually getting those times will involve both CoreBoot and Nouveau.
Absolutely, the nouveau devs are doing an excellent job, no argument here.Radeon driver was already making very good progress and documentation has certainly helped but as Nouveau driver has proved, with enough focussed effort, we can jump through the hurdles necessary. Pidgin, Empathy et all support a crap load of protocols, almost all of them entirely reverse engineered as well.
andIn short, reward good behaviour with your money.
Instead of telling consumers "vote with your pocket books" here is a alternative way of doing it (which has been VERY successful for our company which also does opensource). If you have the specs released, instead of saying "consumer pay us to do it" put a hit list out on the features and pay for independents to take on the task. Example:
Users want flaw free accelerated HD playback, put out ads saying such
"$4000 for a working implementation of accelerated x264 playback in the XYZ's brand foss driver, submit code and upon acceptance payment will be made"
If manufacturer's, commercial distro's, etc would do this it would probably speed up development of wanted features. Get those features in there in a timely manner and then people purchasing your products will follow.