Re: License clarification: BSD vs MIT

David Fetter <david(at)fetter(dot)org> writes:
> On Sun, Oct 25, 2009 at 11:34:28PM +0100, Dave Page wrote:
>> It's not a perfect match to MIT, but it is close. We (-core) are
>> already actively working on this issue to find the most appropriate
>> way forward.

> Legally speaking, what are the issues at hand here?

None really: the Postgres license is what it is. This discussion is
just about what is the simplest description of it. Red Hat has decided
that it fits in their "MIT" pigeonhole better than it fits in their
"BSD" pigeonhole. If you compare the OSI definitions of these licenses:
http://www.opensource.org/licenses/bsd-license.php
http://www.opensource.org/licenses/mit-license.php
our wording is not a terribly exact match to either, but RH's lawyers
think it's closer to MIT. The key point is that they consider "BSD"
to include the no-endorsement clause and possibly the advertising
clause, and we have neither.

OSI uses the phrase "simplified BSD" to refer to BSD-derived licenses
without either of those two clauses. Red Hat has decided to call them
MIT instead. As best I can tell there's no really important distinction
between simplified BSD terms and MIT terms, so there is not a lot of
point in arguing which one ours is.

If anyone comes up to you and starts asking questions about two-clause
vs three-clause vs four-clause BSD, the answer is that we use the most
liberal variant. Otherwise I'm not sure it matters.

regards, tom lane