[License-review] Approval: BSD + Patent License

[Mark Wielaard]

On Fri, Jan 15, 2016 at 12:34:20PM -0800, Lawrence Rosen wrote:
> Jim Jagielski wrote:
> >> We fully expected that the FSF (and others), based on this and from 
> >> their feedback, would formally state that ALv2 was compatible; you 
> >> can imagine our surprise (and disappointment) when, not long after we 
> >> released it, we were told "nope".
> 
> FSF opinions on license compatibility, to steal McCoy's phrase for this
> purpose, seem "like an exercise in whack-a-mole." Roy Fielding's stated
> opinions on that dispute between FSF and Apache were as usual dead-right.
> FSF's objections to Apache 2.0 came out of left field.  

If you read the whole history then it is clear that the FSF was
(at least according to the public record) pretty consistent. Patent
retaliation clauses are not GPLv2 compatible. That doesn't mean they
are not a good idea, they are. And that is why GPLv3 does contain one.
Based on the same language the FSF suggested the ASLv2 should adopt.
That is also what they said publicly they would do when reviewing the
ASLv2 draft and when the ASLv2 was finalized.

The point is not who is "right" about who is reponsible for this dispute
which is now 10+ years old. Reading the whole history shows some people
might misremember or misunderstood what was said by whom, or simply
assumed the opinion of the other didn't matter. Sadly that happens.
The point is just that at least with respect to patent retaliation
clauses and the GPLv2 the FSF has publicly been very consistent.
So if the point of this license is to be explicitly GPLv2-only
(and not just GPLv3+) compatible, then adding a patent retaliation
clause is not the way to go. Just stick with an explicit patent grant.

> Here's another example: Apache2.0 allows AFL3.0 contributions. GPLv3
> allows Apache2.0 contributions presumably including those AFL3.0 parts.
> But GPLv3 doesn't allow AFL3.0 contributions because that license
> "contains contract provisions."  :-)  You figure it out.

This seems somewhat offtopic for this license review. And it seems that
the simplest explanation is that the presumption is incorrect. But I
think it shows that if your goal is to create a new license that is
compatible with another one, then it is good to talk to the license
steward to make sure your assumptions are correct.

Cheers,

Mark