It doesn't actually say anything. It doesn't say you have permission to do anything, it doesn't say "If you do this I won't sue you", it doesn't license patents, it doesn't promise that as long as you conform to the license that you won't get in trouble in the future (that's the "estoppel" comment in the article [1]).
A normal human being reading the license will get the almost-certainly-correct impression that the author of the code will not come after them for anything they do, especially if it's a small usage. Lawyers must worry about two reasonable scenarios: First, some large usage causing the original licenser to decide to change their mind [2]. To put it in terms most of HN will understand, what if you release some code under the WTFPL license and discover that the Trump administration is using it. Would you be tempted to wedge your way into the legal ambiguity and sue them over it? Your specific answer doesn't matter; you need merely believe that there are plenty of people jumping out of their seats and yelling YES! to make my point, as that is enough to send shivers down a big-company lawyer's spine.
And second, even if you are confident that the original author wouldn't sue, if the rights to the package get transferred to someone else, you have to worry about whether that licenser will change their mind as above. If a new party changes their mind about the license, some of your legal defenses get a lot harder. The bigger you are as a target, the more you have to worry about being a bit, juicy target. A 5-person unfunded startup doesn't have a problem where Google does.
It isn't really a license so much as an assertion that the author won't come after you if you use or modify their code with no legal meaning.
Note that as gratuitous licenses are, irrespective of their language, revocable at will (in US law), basically everything you say applies to any license that is not a contract, it's not special to the WTFPL.
You can revoke a license at will, but I'm not aware of any case where someone took something that was open source and successfully retroactively closed it. You can close your future development but if someone takes the old code under the old license and continues work on it, I can think of no case where that was stopped, or even challenged.
So it seems to me there is a real difference here between licenses that proactively are clearly giving you some sort of rights, even if what those rights are may be up for debate, vs. the WTFPL license, which is still technically probably establishes that the author meant to grant you some stuff, but it is not terribly clear what.
> You can revoke a license at will, but I'm not aware of any case where someone took something that was open source and successfully retroactively closed it.
Revoking a license means that permissions previously granted are no longer applicable to any acts taken in the future, so no action requiring a license under copyright (such as copying) would generally be permitted (promissory estoppel protects, to a limited extent, those who initially relied on a promise that the license would not be revoked before it actually was revoked, but generally only to the extent that they had reasonably expended resources in that reliance and would be suffer unjust loss due the violation of the non-binding promise.)
> You can close your future development but if someone takes the old code under the old license and continues work on it, I can think of no case where that was stopped, or even challenged.
Thats because no one has ever actually revoked an open source license on code in significant use, rather than merely not offering it for future versions. And there's a good reason: most people offering open source licenses, even if they change their mind on a particular work, want open source to keep working, which actually relies very much on the shared belief that gratuitous open source licenses won't be revoked, even though the law of gratuitous licenses clearly allows that.
If it ever happens (and courts don't just completely rewrite the law on gratuitous licenses in defiance of precedent to save open source), the whole illusion of safety comes crashing down.
A normal human being reading the license will get the almost-certainly-correct impression that the author of the code will not come after them for anything they do, especially if it's a small usage. Lawyers must worry about two reasonable scenarios: First, some large usage causing the original licenser to decide to change their mind [2]. To put it in terms most of HN will understand, what if you release some code under the WTFPL license and discover that the Trump administration is using it. Would you be tempted to wedge your way into the legal ambiguity and sue them over it? Your specific answer doesn't matter; you need merely believe that there are plenty of people jumping out of their seats and yelling YES! to make my point, as that is enough to send shivers down a big-company lawyer's spine.
And second, even if you are confident that the original author wouldn't sue, if the rights to the package get transferred to someone else, you have to worry about whether that licenser will change their mind as above. If a new party changes their mind about the license, some of your legal defenses get a lot harder. The bigger you are as a target, the more you have to worry about being a bit, juicy target. A 5-person unfunded startup doesn't have a problem where Google does.
It isn't really a license so much as an assertion that the author won't come after you if you use or modify their code with no legal meaning.
[1]: https://en.wikipedia.org/wiki/Estoppel
[2]: http://dev.hasenj.org/post/3272592502/ibm-and-its-minions