Why FreeBSD not popular on hardware vendors

Tyson Boellstorff perlcat at alltel.net
Thu Dec 11 21:45:21 PST 2008

On Thursday 11 December 2008 19:58:14 Chad Perrin wrote:
> On Thu, Dec 11, 2008 at 05:00:11PM -0800, prad wrote:
> > On Thu, 11 Dec 2008 17:28:13 -0700
> >
> > i don't think that's really what is happening, chad.
> > i think there is just some disagreement as to what is considered an
> > improvement.
> So . . . are you saying that increased support for 3D accelerated
> graphics is not an "improvement", and should therefore not be considered
> a worthy goal?

Not so much considered 'unworthy' as it is a balancing of limited resources. 
If I was a hardware programmer, had unlimited time, beer, and cheese dip, I'd 
add everything just because I could.

It would be cool if there was a way to ensure that all <foo> items would be 
supported. However, even then, high performance video would lag. It is often 
proprietary, and many vendors simply won't publish their specs and need a 
reverse engineer to get any support at all. You can't force them to do it, 
and in the case of an open source OS, they may not want the world+dog to see 
their code for any number of reasons. nVidia is a rare exception, and even 
they are not going to put FreeBSD support at the top of their list. 

Unless you have a job at some video chipset maker, and are of a truly generous 
spirit, willing to risk your job in order to publish drivers, it really 
doesn't matter what priority the powers that be give to video acceleration -- 
we can't ask anyone to risk their job just so <foo> works. If the graphics 
devices themselves are sub-optimal, getting related systems up to a 
razor-sharp performance level is like putting nitro and a supercharger in 
your Lada. You'd have to put it in the back seat, because there's no room in 
the engine compartment for it.

That is also why the high performance fax cards I work with only run on 
windows machines. (that's gotta be about the greatest number of oxymorons in 
one sentence -- my brain had two core dumps just parsing it...)

Long story short, there's room for all types. Enjoy the diversity. Fix what 
you can. Avoid the problems you can. Use the appropriate tools for their best 

More information about the freebsd-questions mailing list