On Sun, 2006-06-18 at 01:33, Craig White wrote:
> The simple fact is that Linux is GPL, has always been and will always
> be. If you want to opt out of GPL and use open source, BSD is most
> likely your best option.
Actually it is fairly ambiguous as to whether kernel modules
must be GPL'd or not. In the early days Linus was quoted as
saying that separate code that used the kernel interfaces did
not fall under the GPL - something I'm quite sure boosted the
early acceptance of Linux since it made the possibility
of vendor-provided drivers seem more likely. More recently
he has been seen waffling about this issue and in any
case his refusal to freeze an API has made it unattractive
for vendors to try to keep up with binary drivers that
have to be rebuilt for every version.
> > But, with less restricted code whatever the original authors
> > want to publish remains free and available regardless of what
> > anyone else does with it. The only net result of the GPL
> > is a reduction in the available choices. Someone can't add
> > a proprietary improvement and sell it for the incremental
> > difference. Which means I don't have the option to buy that.
> ----
> dude - there's even less device drivers available for BSD than for
> Linux. There isn't any 'taint' from GPL license on BSD. That's such a
> huge hole in your argument that it renders your argument pointless.
> ----
My G4 powerbook is starting to show its age, but I don't recall
it missing any device drivers - yet there are reasons to think
it contains some amount of BSD licensed code.
> Your opinions suggest that people are too stupid to realize that they
> are contributing unintentionally to an evil code base...GPL license code
> base.
I don't think a lot of people understand the effect demonstrated
by the RIPEM example. I don't think a lot of people understand
the potential for new patent cases to hopelessly tie up any GPL
code in the same 'work as a whole'.
> Companies and people who contribute to the GPL code base are
> undoubtedly choosing with intent to do so.
Some are, of course.
> At least you're not alone in
> your opinion...Steve Balmer would wholeheartedly agree with you.
The history of objective-C validates that opinion. And is
counted by some as a victory for the GPL. If I were Steve
Balmer I'd make an effort to publicize how that worked out.
> > > Anyone who remembers the BSD TCP/IP code stack in Windows NT knows why
> > > some people prefer GPL license over BSD type licenses...it absolutely
> > > prevents the 'embrace, extend, extinguish' of donated code making its
> > > way into proprietary software.
> >
> > Beg to differ here. Microsoft originally wrote their own code
> > and anyone who had to co-exist on a LAN with those early
> > products (like the win95 version that got the retry timer
> > backwards and increased the rate instead of backing off on
> > congested networks)
> ----
> I specifically said Windows NT (NOT Win95, etc.) what are you
> disagreeing with?
NT was still horribly broken, at least until service pack 6a,
well into win2k's lifetime. Just guessing, but that's probably
when the code was updated to correspond to the bsd reference.
> > And where patents are involved the slowdown will be for the
> > life of the patent - or until someone contributes it to the
> > public.
> ----
> umm...patents and licensing are different topics altogether.
They aren't when one of the licenses says that the 'work
as a whole' must be distributed under that same license
with no additional restrictions. If the patent holder
insists on additional restrictions (which is his right and
would be the usual case), then no part of a 'work as a whole'
using that technology may be under the GPL.
> I see companies like Dell that created dkms which enables them to put
> out a dynamically self-updating device driver that upgrades itself with
> each new kernel release. I see companies like Nvidia putting out drivers
> for their hardware. Things are continually improving.
Nvidia claims that their contracts with the chip vendors prohibits
them from releasing details or a GPL'd driver. If it is possible
to put a license-agnostic layer between components that the
FSF can't threaten with their interpretations of 'work as a whole'
and without a big performance hit, the issue might go away.
> I note without much surprise that Apple doesn't contribute the device
> drivers that they have developed back to the BSD community.
They may have the same problem as Nvidia in terms of contracts
with their hardware suppliers. Or it may just be their choice.
I don't have a problem with companies making their own choices
about what is best for them as long as that company doesn't
have monopoly control over the market.
--
Les Mikesell
lesmikesell@xxxxxxxxx
