Re: testing, testing
On Mon, 12 Jul 1999, Gary Kline wrote:
# If *this* idea ever takes off, a DebianBSD version
# wouldn't have these concerns. Pure BSD code would
# continue its own license; likewise with pure GNU
# code. Anything that was a combo would carry *both*
# notifications and would really be GPL'd. Since the
# Berkeley license seems so open I don't see how it
# could be violated....but then, I'm not a lawyer.
I think they would have these concerns unless the GPL'd bits
are LGPL'd instead. No flamewars please... a bit of clarification
first. You write X and put it under the BSD license. You can
link it against LGPL'd code (as a number of FreeBSD programs
do today). You cannot link X against GPL'd code (the BSD kernel
and a GPL'd device driver is one example) without first placing
the BSD bits under the GPL. Though I think it is possible
to relicense[1] the BSD code under another license, this is
usually frowned upon because some feel it to be violating the
license and others feel that it violates the intent of the license.
I can't speak for FreeBSD -core and IANAL, but this is probably
the biggest reason you don't see GPL'd device drivers in the
FreeBSD kernel. It is not that FreeBSD doesn't want to use
them, or thinks they are no good. In fact, some of them are
very good and would be welcomed with open arms if the "other"
license didn't make it darn near impossible.
It would be great if the GPL vs. BSD war didn't exist. There
are programs scattered everywhere that were written under
one of these two licenses and since rewritten under the other
because someone didn't like the first license. The epitome
of code forks. Split the population along some arbitrary line
and make each of them reinvent each other's work because of
religious views over licenses. :(
But I digress... sorry for the mini rant.
-steve
[1] The sticking point is usually the advertising clause
present on most of the BSD kernel bits. This definitely
doesn't mix well with the GPL.
Reply to: