Fedora Core 6 HUGE problem - Binary drivers.
Robin Laing
Robin.Laing at drdc-rddc.gc.ca
Wed May 30 15:37:20 UTC 2007
Les Mikesell wrote:
> Arthur Pemberton wrote:
>
>>> > When they want a feature in the new kernel. Features I am
>>> personally looking
>>> > forward to is the removal of the limit of arguments to exec and
>>> dmcrypt
>>> > getting write barrier support back.
>>> >
>>> >> where the next one is only months away? If there is an answer to that
>>> >
>>> > Your assuming that there will be an nVidia driver for the next
>>> Fedora release.
>>> > That isn't necessary going to be the case.
>>>
>>> But that's not a big problem until updates stop for the current release.
>>> You can put off installing a new release as long as necessary. A
>>> kernel update within a release that breaks needed drivers is a big
>>> problem.
>>
>> But, you're ignoring the part that the entity doing the kernel upgrade
>> and the entity working on the driver are two separate and independent
>> entities.
>
> No, I'm not ignoring that part. I'm saying that _is_ the problem. If
> some other OS regularly made incompatible changes without coordinating
> the availability of drivers with releases they'd have been dead long ago.
>
But some other OS has support of the manufacturers for their drivers in
a big way. This isn't the same so it isn't a fair comparison. I think
nVidia is doing a great job in comparison to AMD/ATI.
>> Seriously, you don't give Nvidia's devs enough credit, if
>> their higher ups cared, their drivers would always work.
>
> Oh, I'm amazed they have kept trying this long. They have to be insanely
> frustrated by something that claims to be an OS but refuses to define an
> interface for drivers.
>
In a thread way back some time ago, this was discussed. The issue, if I
remember correctly is flexibility to make changes to the API, either for
improvements or security. Of course it could be made easier.
>>> Don't forget fedora's 'other' purpose of evolving to near-verbatim RHEL
>>> releases - which then has the problem of frozen application versions for
>>> the long, long term between those releases. I agree that the current
>>> fedora disto acting as a fast-changing testbed is a necessary evil, but
>>> I wish there were something that used the same packaging and admin
>>> techniques that would make a usable desktop - like an FCx release
>>> usually becomes for a short period near the end of its life.
>>
>> What are you talking about? My Fedora desktop was always usable.
>
> Did you have firewire drives mid FC5? A good 6 months of downtime might
> have changed your mind. The Evolution exchange connector was broken for
> about the same interval after a brief glimpse of a working version. Not
> sure what you've been using.... Some stuff works most of the time.
>
Some of the issues are outside Fedora's control. I read the Evolution
list and see enough comments about issues not being dealt with. This
can be compounded by Fedora/RH making changes to their code to meet the
fears of litigation and lawyers big fees. :)
--
Due to the move to M$ Exchange Server,
anything that is a priority, please phone.
Robin Laing
More information about the users
mailing list