Patent issues get us into funny ground, upon which the IETF and others
tread carefully.

For copyright licenses, we have some notion (not clearly stated that
I've found) of "adequately free not to require user to add
ACCEPTABLE_LICENSES+=foo".  Generally the identity of the copyright
holder is not in question, and the right of the copyright holder to
specify a license is essentially never in question, so licenses can be
evaluated without passing judgement on the validity of claims.

With patents, deciding if some use infringes a patent is less clear,
and what patents exist (and even can exist) varies by jurisdiction.
Then, many patents are issued that shouldn't have been, and everyone
has to choose themselves whether to refrain from practicing the
claimed invention or to ignore it as invalid.  It's this judgement
that IETF as an institution has basically decided not to make.

I have no idea in which countries the MP3 patents are issued.  I'd be
stunned if they were valid in every country, particularly as applied
to software-only implementations.   So that raises the ugly issue of
having per-jurisdiction coding of license restrictions.

As wiz pointed out in the email, there's also the issue of the GPL
patent clause.  (I saw something on debian-legal that indicated LAME
had a modified LGPL with additional patent terms.)


But, certainly whatever info is in the distribution should be pulled
into the license file, so people can make their own decisions.

  LICENSE=       fee-based-commercial-use

This doesn't make sense.  Patents don't distinguish between commercial
and non-commercial use, and I didn't see anything on the web when
looking just now that indicates a royalty-free license is granted for
noncommercial use of encoding.

-- 
        Greg Troxel <gdt@ir.bbn.com>