alt Hacker NewsThe copyright holder (the author) is solely responsible for choosing how they want their work to be distributed, and is not bound by any other sort of constraint. They can choose any license at any time, and change their mind however often, and it whatever direction they want. Any previous licenses used (MIT here) bear no effect whatsoever. There is no license in the world (and cannot be) that would prohibit the copyright owner from changing it. It makes no sense, the license terms only apply to the licensee, not to the licensor.
Of course, the author cannot retroactively change the license of any previously distributed work. Anyone is free to fork off Bear from its last MIT code and do whatever they want with it.
So no, the MIT license does not "explicitly allow to relicense a project at any point" (emphasis mine). The MIT license allows licensees to license their derived work however they see fit, it has no effect on the relicensing by the licensor (the copyright holder).
Yeah, sorry if my terminology was unclear here: by “relicense” I colloquially meant to say “assign a different license to the project that is applicable for any work from that point onwards”.
> Any previous licenses used (MIT here) bear no effect whatsoever. There is no license in the world (and cannot be) that would prohibit the copyright owner from changing it.
I don’t think it’s that simple. The Bear project appears to have accepted external contributions under the original license, so the project is subject to that license as long as those contributions remain.
It may not be a big practical issue in this case, due to the MIT license being quite permissive, but if the project was e.g. GPL-licensed, the maintainer wouldn’t trivially be able to change the license in “whatever direction they want”. (And by “trivial” I mean without e.g. rewriting or discarding the external contributions.)