How not to handle a licensing violation

Posted Apr 12, 2007 15:18 UTC (Thu) by mheily (subscriber, #27123)
Parent article: How not to handle a licensing violation

Clearly, the OpenBSD dev who committed the tainted code was in the wrong and it was within the rights of the bcm43xx devs to request it's removal. Since the bcm43xx devs wanted to make the problem a very public issue, it was also Theo's right to go on their mailing lists and complain. What this incident really exposes is the license schism between the BSD camp and the GPL camp.

It's too bad that the bcm43xx team chose to work under the GPL since reverse engineering a complex piece of hardware is such a difficult thing to do that we need all the free software developers we can get. The end result is a driver that is only useful as a "glue" between the kernel and the hardware, due to the use of magic numbers, undocumented registers, etc.

What is even worse is that the bcm43xx team is openly hostile to the idea of Broadcom using their driver to create a proprietary binary-only Broadcom driver for Linux. What's so terrible about having a a fully functional vendor-supported driver that you can study, reverse-engineer, and use to improve the capability of the free bcm43xx driver? Is nothing better than something?

If the driver were freely licensed under the BSD or ISC license, all of the free operating systems -- and even the commercial ones like Solaris and OS X -- could use it and contribute to it's development. A device driver is not like a normal piece of software, it is really only useful for one thing: talking to a specific class of hardware. There is no reason to lock it away behind a restrictive license.

How not to handle a licensing violation

Posted Apr 12, 2007 19:59 UTC (Thu) by ajross (guest, #4563) [Link]

Unfortunately, you have your facts pretty mixed up. The reverse engineering was not done by the bcm43xx team. That work was done in a "clean room" manner by a separate group of people, who make the resulting specification available to the public at:

http://bcm-specs.sipsolutions.net/

The choice of the GPL was made for the code implementing the driver (as it must be, obviously -- it's distributed as part of the Linux kernel), not for the hardware information required to write it.

As far as I can see, this misunderstanding invalidates every one of the points you have made. Nothing was locked away. All of the reverse-engineering work is available. Everyone is free to create drivers under any license.