OT: nVidia driver [was: Wish list]
Bryan J. Smith <email@example.com>
thebs413 at earthlink.net
Thu Jun 9 16:12:08 UTC 2005
Paul Iadonisi wrote:
> Again, I haven't seen many, if any people arguing about whether or not
> nVidia is designing their cards and drivers to an open standard.
I know, I'm the first one who reminded people. Because a _lot_ of the
arguments that I have seen made are applying attributes of proprietary
standards, which are wholly inappropriate and inapplicable.
> But you're starting to sound like Sun's Jonathan Schwartz. To him,
> and some others in the industry, open standards matter more than
> source code under a Free license.
Now I _never_ made that. Why does everything have to be in black
and white? All I said is you can_not_ apply the same logic of proprietary
standards to proprietary source of open standards. That's all.
Again, it's like being a Libertarian and watching Democrats and Republicans
sometimes. I try to interject a fact that is not applicable to the demonizing
viewpoints of both sides, and what I get is people just taking what I said
and demonizing it the same.
Same deal here. I see Freedomware proponents using Commerceware/
Hostageware arguments, that are totally and wholly inapplicable, against
Standardware. I try to point that out and all of the sudden, people like
yourself are saying I'm like Schwartz saying that Open Standards are all
that matter. No, I'm not saying that at all!
I believe _very_strongly_ in Freedomware. I use it by default. I regularly
defend Debian and Fedora on their strict guidelines, and warn of the
dangers of not doing such -- especially when it comes to indemfication.
E.g., I have a _formal_policy_ in my department that _no_ Knoppix or
other Live CD is allowed in the building _except_ those that have been
verified to be "pure" Debian, Fedora or another distro that does not have
I support projects, even if extremely limitedly with a patch of only a few
lines here or there, as much as I can. And I am very aware of the
dangers of not supporting Freedomware and other community
But I'm also an engineer and a realist, and that means I have to make
solid arguments when it comes to adoption, and those arguments are
100% based on risk mitigation of many factors.
> That's not the case with the Fedora Project and I, at least, am going
> to lobby to keep it that way.
And I will too! If you didn't notice, in the init thread, I was very much
against even looking at Solaris' new init, because of that exact issue.
If an end-user is going to make that decision based on all collective
risk factors, then it's their choice. But from a development and shipping
standpoint, I agree.
I think the problem is that once someone started calling anyone who
recommends nVidia's driver a "zealot," that's when I thought it was
rather hypocritical. And that's when I came into it. Maybe I should
not have, especially since this was the "development" list.
> And although that doesn't mean deliberately breaking closed source
> kernel modules, it does mean having zero concern about whether or
> not they break. We leave that completely in the hands of those who
> have chosen keep their source closed.
Agreed. And I can understand your viewpoint of expected support.
Anyone who chooses nVidia's Standardware drivers must contact
nVidia for support, and do their "own homework" on configuration
management. I.e., people who think "upgrade to the latest" is
a valid configuration management are the types of people I _flog_
in person (at least virtually ;-)!
> But I know my folly, and that it is *my* folly. I help people to get
> it working when the latest kernel update precipitates a another
> breakage. Never, NEVER, would I argue that Fedora should make it easy
> for VMware to keep their drivers closed. I want pressure to be placed
> on VMware to open at least those kernel modules so that core kernel
> developers can help when they break. EVEN when it inconveniences me.
And I understand your viewpoint, and the collective frustration, on the
nVidia issue. I just didn't like the "over-extended comments" made by
a few people. One must be cautious when making such.
Bryan J. Smith mailto:b.j.smith at ieee.org
More information about the devel