Fri, 06 Apr 2007
OpenBSD in a GPL Violation
Remember how Greg Kroah-Hartmann got flamed by the OpenBSD people for offering to write a GPL-licensed driver for everything, based on an assumption that open specifications are better for general public than the open (GPL-licensed, in this case) driver? Well, my feeling always was that you need an overly big ego to create and manage a fork of a big project, like Theo de Raadt[?] did with OpenBSD.
Now guess what happens when somebody accidentally finds out that his GPL-licensed code has been committed to OpenBSD without even asking for relicensing under the BSD license. This has happened even though there are unrestricted specifications available for the hardware in question.
I find the whole discussion in the linux-wireless list
to be quite enlightening. Despite the bcm43xx developers
clearly stating that they are willing to relicense the parts of the driver
under the BSD license, provided that OpenBSD developers really ask
for relicensing, the Theo's resolution of the problem was to simply
delete the bcw driver in question and then to
moan
about Linux bcm43xx developers being evil and against a public good(tm). Which is the worst possible solution for both OpenBSD team and the OpenBSD users.
Also the interesting thing is that this case clearly shows the superiority of GPL over the BSD license in its fairness which is well balanced between the developers and the users under the GPL, unlike the BSD license.
Disclaimer: I own a laptop with a Broadcom 4306 wireless card
and I think it was quite a heroic effort on the bcm43xx people side,
to split their team into two parts which never communicate with each other
directly, one part writing a publicly available specs based on reverse
engineering the existing drivers and firmware, and the other part writing
a Linux driver as a clean room
design based on these specs.