[geofft]

The Apache license has an explicit patent grant (the MIT license says "permission to use", which isn't a copyright grant, so it probably has an implicit patent grant), and an explicit statement that patches intentionally submitted for merge are submitted back under the Apache license.

The reason we have licenses at all instead of the Unlicense or similar is to make things unambiguous for courts and lawyers. Explicit is better than implicit. The length of the MIT license isn't actually a feature.

(And the contribution section seems like it avoids licensing issues that MIT doesn't.)

[dmm]

License bikeshedding is fun, so please indulge me.

People dislike apache because it's complicated and requires annoying notices on distribution of modified versions. Debian and fsf say it's free. OpenBSD believes the patent provisions are non-free and refuses to include apache licensed software.

I think a project is better off having non-trivial contributers sign explicit license grants, even you admit that explicit is better than implicit.

[geofft]

I don't understand how the OpenBSD project defines "free", so I can't usefully comment on how they consider the Apache license "non-free".

It sounds from https://softwareengineering.stackexchange.com/questions/2632... like they are reading the existence of an explicit patent clause in the Apache license (regardless of what that clause is!) as an "additional restriction". I want to know whether they believe the MIT license has a patent grant, and if not, what they think "Permission to use" means.

[liveoneggs]

until they are taken to court the licenses are all just assumptions. Did OpenBSD have a lawyer review this decision and post the language somewhere?