I wasn't actually referring to bridgman, but she felt she had to pipe up. Well I'm all for greater inclusivity and free speech, so please feel free, my dear.
Then we have the issue of "turning the tables". Suddenly, I am a defendant who is obliged to answer spurious calls for evidence. I'm not the one who has paid out MILLIONS in out-of-court settlements to complainants.
Well here's the thing : the Linux drivers both for nvidia & ATI are an absolute disgrace. They are *so* much worse than their Windows counterparts that only a muppet or a paid apologist would find this acceptable.
Let's have an explanation for that, guys. Instead of fulminating and posturing, let's have the full details as to why the Linux drivers are so awful. When you do that to my satisfaction, then I might consider commenting further.
My apologies. You referred to people asking for citations in the plural, and only two of us asked, so it seemed that I was one of the stooges under discussion. I put RealNC in the <ad hominem> bucket, hope I got that right
re: turning the tables, you stated that something none of us had heard of was "a matter of public record"; it seems reasonable to me that we would ask for specifics. Why do you feel that is inappropriate ?
As far as the relative driver quality between Windows and Linux, the core issue is the resources we are all able to allocate relative to the needs of the current Linux graphics environment. The best market share numbers we can find all indicate that Linux non-server market share is just under 1 percent, compared with 8-ish percent for MacOS and 90-ish percent for Windows. We allocate relatively more resources to Linux than the market share justifies, and we all try to leverage common code across multiple OSes where possible, but that still doesn't mean Linux gets the same resources as Windows.
If it were easier to write Linux drivers than Windows drivers that would certainly help the Linux experience, but if anything Linux is actually *more* difficult to support than Windows in most ways. The key factors seem to be (a) the large number of slightly different distributions we need to support, (b) the ability (and tendency) for users to mix and match subsystem versions, and(c) the fact that the graphics subsystems are constantly evolving as opposed to the "new release every few years" model used by Windows.
With a tiny fraction of the market share and a relatively more challenging environment than Windows I think everyone involved will admit that providing consistently satisfying Linux drivers is hard work. The good news is that the massive changes of the last year or two should start to settle down by the end of 2009, and hopefully we will all be able to deliver an excellent user experience from that point on.
Our hope is that by actively supporting open source driver development we can give you a better experience with the newest X servers and kernels, while the proprietary driver gives you performance and features closer to what you would see on other OSes, on a "supported" subset of distributions and hardware.
Thanks, I am grateful for a carefully considered, well thought out reply. You neatly avoided the obvious trap of "either you are incompetent, or the management are deliberately holding you back. Either way, you are guilty as charged". And you are quite right, the pesky users are partly to blame ...
I take it that we can expect driver expenditure and resources to increase as the Linux market share overtakes Windows. It will be nice to get the full benefit of the synergistic heart of ATI drivers, which comprise a holistic quintessence that redefines the interactive boundaries of the object-oriented user experience, while leveraging an admixture of legacy and bespoke paradigms, to enhance the inclusivity and response of the user-oriented workflow.
Let's be fair. Neither ATI nor NVIDIA have consistently released good drivers for Windows either. If Microsoft ARE in fact colluding with them they are getting a pretty shitty deal.
I agree with bridgman in that the state of X or the graphics stack so to speak has alot to do with driver quality. Our "next-gen" graphics stack is yet to settle. The TTM/GEM saga is an excellent demonstration of that fact. Or look at hardware video decoding, with a number of APIs and proposed solutions. At the moment alot of dev work is going into getting the infrastructure "right". Once that's done I'd expect drivers to get more love.