ATI video comes out of the closet

[Les Mikesell]

Jose Celestino wrote:

>>> Both Linux and Windows work on platforms that have, literally,
>>> thousands of vendors manufacturing a tremendous range of equipment,
>>> most of which has to have a properly working device driver.
>>  Yes, and my experience over the last 5 years has been that the Windows versions are more dependable than the fedora versions.  I'm sure there are 
>>  individual exceptions to that, but I just don't see fedora as a bastion of stability here - or in a position to claim that they have the only 
>>  approach to drivers that can work.
>>
> 
> What? You must be trolling. More dependable? More like more predictable,
> you can always predict there will be troubles.
> 
> And if your experience is from the last 5 years I bet you've had ME and
> 98 ubber troubles to some extend.

Yes, I could never get a windows box to run more than a few weeks with 
some applications back in the the days before NTsp6a and win2ksp2, but I 
don't see how that is relevant to a current discussion.  There are lots 
of ways to do things wrong, but once they are done right all you have to 
do is stop changing them.

>>  > Much as
>>> they'd like to, Microsoft can't control all these vendors; the original
>>> PC was wide open--they even published schematics and the source to the
>>> BIOS--and that legacy is embedded in the attitude of the vendors today.
>>> (MS's attempt to lock down the driver interface with Vista is meeting
>>> with a lot of resistance.)
>>
>>  The Vista approach deserves to fail for the same reasons DRM does, but the driving force has to be consumer reaction.  If something is difficult to 
>>  use, don't use it.
>>
> 
> What are talking about? Is it dificult to install or use Fedora? What's
> the dificulty, I don't get it. Could you elaborate?

On the machines where I've used fedora, things that worked in the 
initial install have broken regularly in updates.  These include 
firewire drives, mainstream scsi controllers used by IBM and DELL, and 
something on an IBM server that I haven't figured out yet where at one 
update version required a bios flash update after which the previous 
kernel wouldn't run.  Most of these problems happened mid-version - and 
the list archive is filled with similar issues - and this is even 
without considering the times kernel updates have been pushed before a 
corresponding binary driver like nVidia or VMware is available.  These 
are specifically fedora issues, though, not Linux in general.  I have, 
for example, an RH7.3 box that has been up for 4 years (no reboots) and 
a large assortment of centos 3.x boxes that have been up for years 
except for reboots following kernel updates - I'd consider these 
statistically equal to current non-Vista windows versions in not 
crashing due to driver bugs, but note that is with a 2.4 kernel.

-- 
   Les Mikesell
    lesmikesell at gmail.com