> Linux *IS* a commercial *NIX (this also applies to *BSD btw)in several > varieties of its distributions with excellent support. Better and quicker > support than RedHat's corporate offerings? I don't know about that, > perhaps you have some means of backing up that statement? > Ummm, forgot about their support desks... (but this does *not* make Linux commercial!) > > machines ain't meant to run games and if their intended to proceed > > sophisticated 3D graphics for simulation and the like they'll highly > > likely have professional high end GL boards installed that have > > Point taken. But many people doing high-end 3D graphics are scientists > that like to run Free OS's because they're that type of person and also > because they use them for the floating-point intensive number-crunching in > their work. So, one woould need to know how large/small an audience that > is before writing them off as negligible. I don't have any data on that, > do you? > All I can say is, I've been working at a research institute at the local university (which doesn't have too many funds to spend anyway) for three years and most of the scientist around used M$ for office stuff and fat SiliconGraphics for their main work (that Onyx did a damage worth of a Ferrari back then when they purchased it...) I was the only guy who tried Linux (on a 486 btw :) ) and no one noticed at all, their central point usually was "I've got to put it there on my desk, switch it on, work with it, not fuddle around with man pages, IRC chats and upgrade to bugfixed versions all day long, that's not what I'm getting paid for." > > commercial drivers shipped with card. So why waste money on setting > > aside a coder for programming drivers for a handful of guys who wanna > > play Tux Racer on their machines. The better place to start IMO are the > > Depends on how large the hobbyist home-user market is in reality. If > there are a couple of hundred thousand in that "handful of guys" and they > have to decide "What do I want as a decent GPU in my dual-boot, nvidia or > ATI?" then ATI can steal a march by dedicating $60K to a coder for a year. Then this forecast must be reliable - those couple of 100K users really *have* to migrate to ATI, and drivers is not the sole secret of success, there's performance, support, features. Hmm, well - ATI card's features usually top NVidia anyway. Ok. Anyway, how many bucks does ATI make on selling one board? design cost, manufacturing, managers, product placement, advertising and all the words that come to my head but I'm too lazy to look up in the german-english dict that have to do with costs of running a company? A dollar? 2? maybe 10? Let's say 10? What if a linux supported board draws away only half of the folks from Matrox/NVidea/whoever? Comes down to 10x200.000 user= 2meg$. Still not worth the effort. > > > Hmyeeessss but this is against the idea of an free and open OS, isnt't > > it - solutions like this rather apply to the above mentioned companies, > > I think. > > Yes, but the original poster has a card, wants to run XFree86 (I don't > know if he's *BSD or *NIX or whatever) so that he can run his underlying > OS and this is an unpleasant but pragmatic solution. Ideally he would sit > down with the source and code-up a driver in a sweaty, 97-hour, > caffeine-fuelled hackathon and GPL it to great applause, but this is the > Real World (TM). I'd do if only my C / OS design knowledge would suffice... :) So, again, advice to the original poster: either don't > use the card (the morally pure option which may also work out cheaper) It won't. Would have to buy another card and will only get half list price for the current... > And in general: if planning on installing a Free OS of some sort then > check out the supported hardware before buying an expensive new item. "Yes! Yes! To Obi-Wan's words listen!" :) Dex
