[laeri]

The author didn't seem to request payment in monetary form but expected some kind of contributions back which would have helped both sides. It would probably be difficult to include some guarantees about upstream contributions into the license but interesting takeaway.

[ahepp]

Doesn't that bring us right back to GPL family licenses?

[pabs3]

The GPL doesn't require contributing back, only contributing forward to users.

[kweingar]

With GPL you don't have to actively work to upstream your patches, but in practice you can't withhold your patches from upstream. If you add a feature, they get to have it too.

Unlike permissively licensed software, where you can add proprietary features.

[pabs3]

Depends how savvy your users are, and what your users lose if they do send your patches upstream. For example, GRSec or RedHat both drop you as a customer (so no security updates) if you republish their patches publicly. Or a paid iPhone app's users probably wouldn't know what source code is, let alone where/how/bother to republish it for the benefit of other users.

[ahepp]

It seems pretty difficult to legally hide useful code from a GPL upstream.

But, if the argument is that the GPL is too permissive to achieve what the author wants, why on earth was the author using the MIT license?