Opinion: NVIDIA drivers are a BAD Thing [tm]

Sean Estabrooks seanlkml at rogers.com
Tue May 18 18:20:24 UTC 2004


On Tue, 18 May 2004 11:36:53 -0600
"Rodolfo J. Paiz" <rpaiz at simpaticus.com> wrote:

> At 01:56 5/18/2004, Sean Estabrooks wrote:
> >If we can discard the argument about whether open source is preferable
> >to closed source things become simpler.
> 
> I think we can; certain *I* can.
> 
> >Being simplistic then, the more "votes" that companies
> >get because they support open source (intel seems to be a pretty good actor
> >these days for example)  the better.    The fewer "votes" that companies
> >who don't support open source get (NVIDIA in this case) the better.
> 
> We can agree on this too.
> 
> >Everyone has to decide for themselves if their immediate needs outweigh
> >the benefits of  not rewarding closed source companies.   If *everyone* was
> >really idealistic and joined forces supporting open source companies there
> >is every reason to believe the market would respond favourably.  However,
> >it seems more people are content to just use whatever works and not follow
> >a more principled course of action.  Such is life.
> 
> I hope this part is not directed at me, since my point here is:
> 

Nope.

>          1. I got a nice machine (Athlon XP 2200+, 256MB, 40GB, eth, modem, 
> cd, GeForce MX 400, 19" LG 900B monitor, UPS) for $400 from a friend whose 
> business went broke. No one asked me what video card I wanted, and spending 
> another $100 or more on another card was not an option, which is why I was 
> delighted to get this great system for cheap in the first place.

You had to decide for yourself if the immediate cost saving
outweighed the downside of going with a closed-source provider.


>          2. For me, the nv driver sucked. Four hours of effort, and I had a 
> horrible 800x600 resolution where opening a terminal window would cause 
> huge amounts of white noise and flickering lines to appear on the screen to 
> the right of it. Miserable, just miserable. (And not necessarily the 
> general case... just what happened on *my* machine.)

Yeah.   The nv driver suffers because of lack of documentation
about the cards it tries to support.

>          3. The nvidia driver worked and did so *flawlessly*. The 
> installation was easy. It is an acceptable solution for now.
>
>          4. I will now happily use this *working* system to do work, 
> including promote Linux as fully-functional and as-good-or-better than 
> Windows for my office. I will use it in many other ways beneficial to my 
> happiness, my pocketbook, and the Open Source movement.

Ok.   but you don't disagree with the underlying economic principles do
you?   Perhaps it will all work out.   But in an ideal situation you would
get to promote the open source movement with open source software
instead of closed source software right ?

>          5. I will also use this working system to test and debug the 
> open-source nv driver, which I could not do if I (a) didn't have the card 
> at all or (b) couldn't use the system for anything other than to test the 
> driver. (a) would be nice but reality happens to disagree, and if (b) then 
> I'd have to sell my three-month-old system and get laughed at by my boss, 
> who would take it as proof that Linux is a shackled, hobbyist operating 
> system which cannot be reliably deployed across an enterprise.

Sure.   but a better sollution would be if the hardware manufacturer
provided a open source driver and that's what my comments were
about.   Purchasing hardware from them anyway, rewards them
for closed source offerings.

>          6. I have already complained loudly to NVIDIA about their drivers 
> being closed. I have made it clear that this *will* be major factor in 
> future purchases.

Nothing is louder than dollar votes :o)

>          7. In the future, we all want open-source. *NOW*, we want 
> open-source *and* we want more widespread adoption of Linux. For users who 
> are converting from Linux, being told to throw away a $100 video card 
> because its unsupported sucks and reflects badly on Linux. Being told to 
> throw away the same $100 video card which has perfectly good working 
> drivers from the manufacturer because those drivers are not open makes us 
> look like idiots. In marketing, customer perception *is* reality.

This is something I hadn't considered and is a good point.   Wish more
of us still voted differently though.

>          8. I've already made at least $50,000 worth of buying decisions in 
> which preference for open competition (AMD), preference for open standards 
> and competition (StarOffice), and preference for Open Source (Linux) were 
> important factors. When I actually buy a video card, rest assured that the 
> availability of good open-source drivers will be a major factor in my brand 
> selection as it has been in the past.
> 
>          9. But I will not, repeat will not, alienate the business 
> community around me which could mean converting another 50 happy users by 
> doing something which is damaging to me in the short run, totally 
> unnecessary in the practical sense, and asinine in the views of my peers 
> and bosses. I will not win a battle and lose the war. And I _will_ 
> challenge simple-minded, uncompromising black-and-white views such as Rui's 
> when presented as the One True View. I'm fine with people choosing one side 
> of an argument, but it bothers me when they seem incapable of seeing both 
> sides... zealotry is not a good thing for Linux.

Not suggesting zealotry.   Just talking about leveraging market principles for
something we both agree is a desired outcome.

Cheers,
Sean.





More information about the fedora-list mailing list