To counter this argument, I would point out that I don't normally purchase used 3D acceleration hardware, and that by the time these cards are "old" they will also be "obsolete," meaning that you will have sunk a good amount of money into hardware that didn't work properly for you until it was outdated.
nVidia's drivers, on the other hand, have worked with brand new hardware since I owned a Riva TNT (1998). They continue to work. Anybody with a new ATI card, however, has to choose a manner in which they are going to cripple X.Org. Do you want composite or DRI? Before, the choice wasn't even an option. Also, fglrxconfig output bad xorg.conf files last I checked, because I had to hand-tune mine to get the server working. Justin On 7/1/05, Jens Mayer <[EMAIL PROTECTED]> wrote: > * On Friday 01 July 2005 01:49, Justin Hart wrote: > > Buying an ATI card for a Linux box is not a good decision. Go with > > nVidia, at least their drivers work. I've thought of buying an nVidia > > card for this notebook for months because, frankly, ATI hasn't been > > taking care of the matter, and won't in the forseeable future. > > There's two sides to every story. It is true that nVidia's drivers are ahead > of ATI's counterpart, especially on desktop computers. They seem to be much > more stable and mature, even while adding new features more quickly, i.e. > support for Xorg's render and composite extensions. > > The situation concerning notebook-specific features is a bit like playing > roulette. Most people want to use suspend to disk or suspend to RAM on their > quite expensive laptops, and it's both drivers who often fail miserably in > that case, whether they are from nVidia or ATI. There are known workarounds > which might or might not get the stuff working, the chance of failure is > high, depending on numerous other things like the driver for your > framebuffered console and so on... Guess what? The open source drivers > usually work, but do not offer 3D acceleration in many (ATI) or all (nVidia) > cases. > > Which brings us to another important point: Contrary to nVidia's practice, ATI > gives the specifications of older cards to the developer community. That's > why there is an open source alternative for ATI's cards up to and includig > the Radeon 9200 with working 3D acceleration support, and that's simply why > there is no real open source alternative for nVidia cards if you want to use > 3D applications on your box. > > Not that important? Well, while the ATI Mobility FireGL T2 in my IBM laptop is > not yet supported by open source drivers, it certainly will be in the future. > I wonder who's first in offering a 3D accelerated driver really supporting > suspend to disk on my laptop: ATI or the guys from r300.sf.net. ;-) > > Now vice versa: The Geforce2 GTS in my desktop is quite ancient, but was good > enough to play around with Xorg's composite and render extensions to get some > solid eyecandy. Guess what? nVidia decided to not support those cards > anymore, they now just get the most important bugfixes via some (yet to come) > "legacy drivers". Now that means a very little chance to have the new and > still experimental stuff getting developed in my card's drivers in the > future. Open source alternatives? None. See above. > > "Buying an ATI card for a Linux box is not a good decision." is too general to > be answered with "yes" or "no". > > Regards, > Jens > > -- > Reporter: "What would you do if you found a million dollars?" > Yogi Berra: "If the guy was poor, I would give it back." > -- > gentoo-user@gentoo.org mailing list > > -- Justin W. Hart -- gentoo-user@gentoo.org mailing list
