An explanation of the difficulty of solving licenseproliferation in one sentence

Alex Bligh alex at
Wed Mar 9 19:42:41 UTC 2005

--On 09 March 2005 14:33 -0500 Ian Lance Taylor <ian at> wrote:

>> Finally, mixing is almost never required. Most creators of
>> derivatives have no need to mix code under two different licenses
>> because they create adaptations which will retain the license.
>> So we are arguing over what?  Less than 1% of the reuse cases?
> I really could not disagree more.  I have seen many cases of attempts
> to mix code with incompatible licenses.  The most obvious ones are BSD
> people rewriting GPL library functions for GNU/Linux compatibility.

Yep. But isn't much GPL code that gets rewritten this licensed under the
GPL (as opposed to say BSD) precisely BECAUSE the authors didn't want them
to be compatible with BSD licenses and thus able to be incorporated into
works without reciprocal licenses?

If this isn't the case, can we assume the FSF (to whom much GPL stuff has
been assigned) will be dual licensing it all under a BSD license? I think

The point here is the license is the AUTHOR's choice. Good licenses will
thrive the same as good code thrives (viz. ESR's bazaar analogy), as if I,
HP, Joe Blow, whoever, doesn't like the license, doesn't like the cost of
evaluating the license, can't read the license because it's written in
English/French/Klingon, then they have the choice of not using it, and
using something else instead. That's "the market" / "natural selection".
Same as closed-source, except of course they are guaranteed a full refund
on FOSS :-)


More information about the License-discuss mailing list