Hey folks, I was actually one of those participating in this survey, and to be honest, the results are amazing.
I put several answers where I could (Most areas, no?) because it is very important that Linux got whatever it could and it doesn't matter if KMS or DRI2 should have a priority, they both should be in, because they both increase the appeal of using Linux as a better operating system.
Regarding my points of interests especially though, I'd say that Gallium3D is actually something I am going to look forward to, as well as Nouveau in it, because when Nouveau gets up to par with the binary driver, we won't need to be dependent on the NVidia drivers. The composite 2d/3d performance will always be better, bugs will be fixed almost as soon as they are reported. And besides, 1 Infrastructure for all drivers, why have 10000000 different drivers for different GPUs?
But seriously, Linux users, congrats, I think with this survey we have a chance to impress even the big companies who think we're a minority (And yes I'm sorry that I'm using Vista right now but gaming and music creation is something I need! Don't worry though I use Linux for those stuff too just that there's just not the same amount of content)
How can people keep saying linux is not suited for "normal" users if 56% of it's users describe themselves as "mainstream users"??
I think most people read it as "Mainstream Linux user" instead of "Mainstream computer user." Mainstream Linux users generally do crazy things like compile software. Not mainstream in the grand scheme but mainstream for Linux.
What's also interesting is the number of folks that find it necessary to edit xorg.conf manually. I think that alone should serve very well to say that configuring X is a pain and probably impossible for grandma. Well... that is if grandma plays games and wants 3D. Most distributions are getting good at reasonable defaults that "just work." They work well enough that unless you try to game you won't notice. Lots of grandmas don't game, don't use multiple monitors, etc.
Ether way you look at it, I think there's plenty of room for a configuration tool which is needed toward the end of Linux adoption by the average joe Windows (mainstream) user.
I was disappointed that Desktop Effects beat out gaming if by a smidgen. That's just me I guess. On the other hand I'd care more for desktop effects if they didn't break my gaming, and if it didn't require so much tweaking and configuration. At home I spend more time running full screen games than I do wobbling windows or spinning a cube. At work, I don't need to impress anyone or spend time setting up desktop effects. So for both work and play it just doesn't fit for me. So I voted for games games games!
The other interesting thing that surprised me was the adoption of X11 release 7.4. I expected more 7.3s and 7.2s. I suppose those results could be biased by the type of people who frequent here at Phoronix which seem to be more the bleeding edge type crowd. I'd be in the 7.4 group but something with my distro repository and the proprietary Nvidia drivers has a problem with an x11-input package. The 3d stuff works but it kills my keyboard and mouse. If I stuck with the distro provided drivers it'd work but I wouldn't be able to try the latest beta drivers out. Still, I'm glad to see that 7.4 is out there so soon.
What bothered me slightly was the lack of an option to vote for XRandR. As a laptop user, it is more interesting to me than pretty much anything else bar stability, as projectors and external monitors are an important part of using a mobile laptop in any kind of business environment.
In my mind, xrandr was in "Display-related features". Considering it's mostly waiting for all proprietary drivers to pick up, it's more in "What do you expect in proprietary drivers" than "What are you most interested in Xorg".
Whatever xrandr does, as long as the drivers don't support it there's not much point. And seeing how users don't really use the graphic card tool to configure their setup, you'd say it would make sense for the drivers publisher to implement xrandr support a bit faster.
That's not my theory at all. It's fact that despite best intentions and planning thing don't always go as well as expected. Until actual implementation is actually out there and performing as well as expected it's pure speculation.
What are you looking for in it? If the system can automagically deal with having a projector plugged in (this requires a reasonable amount of magic) it makes selling Linux as a laptop OS to business people who expect to make presentations a lot easier. If the projector has to be configured by tweaking xorg.conf, and restarting, you do not have a business laptop OS.
That magic is XRandR, it may need driver help, but to say it is just a matter for the proprietry drivers to catch up is nonsense as, IIRC, openChrome is an open source driver.
To push Linux on laptops, you needs automagic and effortless multi-monitor support, you need hot-plugable displays, and you need all this to be completely invisible to the user. Windows has done it for years, as has MacOS. Linux has "innovated" with eye-candy instead...