Cc: Jacob Meuser <jakemsr@...>, Theodore Tso <tytso@...>, Adrian Bunk <bunk@...>, Can E. Acar <can.acar@...>, <misc@...>, <linux-kernel@...>, Daniel Hazelton <dhazelton@...>, Eben Moglen <moglen@...>, Lawrence Lessig <lessig_from_web@...>, Bradley M. Kuhn <bkuhn@...>, Matt Norwood <norwood@...>
There are two highly relevant angles to this that nobody is mentioning:
1) Does it make sense to share code, at a technical level?
The fact is, BSD and Linux wireless stacks are quite different. Linux
also has a technical requirement that "Linux drivers look like Linux
drivers." This enables a vast array of source code checking tools like
Coverity and sparse, as well as maximizing human reviewer bandwidth.
Therefore, there is a strong /technical/ motivation for the source code
to diverge. That's quite natural.
2) Information sharing is both rampant and healthy.
Linux and BSD projects share a vast amount of hardware knowledge,
information on how to properly program hardware. Linux folks use BSD
code as /reference documentation/, and BSD folks do the same with Linux
code.
This is far more efficient in many cases, due to the natural divergence
of the respective codebases. It is often easier to look at codebase A,
and then mentally translate that into code for codebase B, than to
directly copy and reuse code.
Jeff
-
