Future of my support for ATI

[Dieter]

me262> Windows is just too insecure and unstable for the hard,
me262> rigorous requirements of Offense/Defense government agencies
me262> (DOD for example),

True, but that doesn't stop the idiots from trying:

http://www.slothmud.org/~hayward/mic_humor/nt_navy.html

(That's just the first hit from google, there's plenty more.)

-------------------

bridgman> Only one person asked if I got her phone number.

Phone number? How quant. Did you get her email address?

bridgman> Everyone else asked what distro she ran at home.

So what OS does she run at home? (The best OSes don't have
"distros".)

-------------------

Dragonlord> Furthermore do you count in the millions of servers
Dragonlord> running Un*x? They do not produce "Browser Hits"...
Dragonlord> they get 'em

rbmorse> They don't buy video cards, either.

Apparently some servers do use GPUs. I forget what they were
doing with them, maybe GPGPU.

-------------------

bridgman> Regarding video playback, have you tried the open source
bridgman> drivers on your 1650 recently ? Alex put a lot of work into
bridgman> adding good EXA render support and Textured Video (which
bridgman> works nicely even under Compiz) and that is now available
bridgman> on both the radeonhd and radeon drivers. The only missing
bridgman> thing in the framework right now is a good vsync solution
bridgman> in the compositor but that is being worked on.

XvMC?

BTW, is there any hope for sync-on-green support?

[deanjo]

Round up of all desktop marketshare stats can be found here along with the growth rates.

http://en.wikipedia.org/wiki/Usage_share_of_desktop_operating_systems

[energyman]

yeah, but that are still browser stats, aren't they? (also wikipedia...). I know that I had to set my brother to IE on windows for some sites and usually forgot to set it back afterwards for a while...

[deanjo]

yeah, but that are still browser stats, aren't they? (also wikipedia...). I know that I had to set my brother to IE on windows for some sites and usually forgot to set it back afterwards for a while...

Sure it's wikipedia, but it does have all the corresponding links. As far as browsers go, the effect of browser agents is minimal. You could also argue the fact that how many of those linux clients are running in a VM on windows.

[sundown]

All in all the linux statistics look shitty. Scarry how bad they actually look!

[Dragonlord]

Whoever had lectures on PES ( Probability and Empirical Statistics ) knows that you can proof anything and nothing with statistics. You just have to ask the right people the right questions. Hence those stats are as insightful as a bag full of poo.

Bottem line is that Linux is far more used than people think. I see this on how often people ask to use Linux or have a Linux specific question.

[deanjo]

I see this on how often people ask to use Linux or have a Linux specific question.

And this is going by the questions posted where? As far as people going "Well I know XXX amount of people running XXX OS" proves nothing as people who tend to have similar views and points tend to associate with people that have the same views and points as well. Browser stats are probably the most reliable source we have as people generally don't "surf the net" on a OS that is not their primary OS. Fanboy site stat counters are useless as well as again, people that go to those sites are generally the ones using XXX OS.

[Dragonlord]

Browser hits can be generated by anything from a user down to bots. Doesn't sound very representative to me.

[duby229]

None of those stats have any value, first of all they simply cannot account for the fact that most linux users are smart enough to run with no-script, or another similar protection in place. Also it doesnt accout for any sort of proxy or firewall... I'm sorry but any statistics gathered by page hits is flawed.

I'd be willing to bet that a --hell-- of alot more linux users run no-script then windows users, and additionally a far greater percentage will be running a properly configured firewall, or proxy. which makes OS identification impossible. The perfect example of the point I'm trying to make can be proved with nmap. Use nmap to try and identify what OS your running, provided that you have a properly configured firewall it cant do it.

Not to mention that a good chunk of the linux market are running on thin clients in an office somewhere with no web access.

[energyman]

@deanjo well I don't search for the people. If someone moves in he comes to me to get internet access (or one of my two colleagues). And i ask which OS they use to help them setting it up. And almost all are Xp, some Vista, very few linux, only one MacOSX so far. In years