For approval: MXM Public license

[Matthew Flaschen]

John Cowan wrote:
> I'm very skeptical of this talk of implicit patent licenses.

Implied/implicit patent licenses are well-documented.  See
http://www.wilmerhale.com/publications/whPubsDetail.aspx?publication=2326
and http://www.patentlyo.com/patent/2008/04/implied-license.html for
just a couple good discussions.

  For years
> we assumed that the MIT license had an IPL because of the word "use",
> and along comes MIT and disclaims that absolutely.

They don't have any right to disclaim the text of their license.

> I don't think you can produce any language in the OSD that requires a
> patent license as a condition of being open source.  OSD #3, per contra,
> clearly requires a copyright license through the term "modifications
> and derived works".

It says nothing about copyright.  It says what the user must be able to
do, period.

>> Thus, I think this license is non-approvable in its own right, and very
>> similar to unsuccessful past licenses such as Broad Institute Public License
> 
> The BIPL tried to *retroactively* exclude patent rights that the
> license had granted, on the grounds that MIT couldn't keep track of what
> conflicting exclusive rights it might have already granted.

There was nothing solely retroactive about it.  2.1 of Broad was almost
exactly the same.  It removed the explicit patent license, while
continuing to disclaim an implicit license.

> In a world of submarine patents, *all* programmers must be prudent
> about the risk of exercising the rights laid out in the OSD or anywhere
> else.

Agreed, but an open source licensee should not fear being sued directly
by the licensor, over patent claims that are essential to any use of the
program.

Matt Flaschen