Hey, thanks for sharing this it’s very interesting.
I’ve looked at you code in details (wish I could do it earlier sorry for the late reply), and I can really see usage for FP on NamedType.
For instance, I suppose a functor would be useful, like in the example you put up, where you triple a number. You can’t pass a strong type into a function double -> double without definitively losing the strong typing, even with the FunctionCallable type. So a functor would allow this new feature. We could add a .map method on NamedType to implement that.
As for monad, I understand how the bind method can be implemented like you showed. Can you think of a concrete example where it would apply? Also, strictly speaking, shouldn’t a monad have a “pure” function that wraps an A into a Monad A? Although I guess the bind could be useful even without that.
Would be interested in your opinion on all that! 🙂

Fair enough, I removed the standard iterators from the comparison in the post. Thanks Daniel!

I agree that operator-> is the right choice here, and with your example of std::optional as prior art, but iterators do model pointers. Unless they’re RandomAccessIterators they don’t model *all aspects* of a pointer, but they model the parts they can. In particular, an iterator usually does not *contain* the object it points to, unlike std::optional and these strong typedefs.

I agree with the post: not entirely, but somewhat. I think you should think very carefully before doing this, but sometimes you want only semistrong typing. It’s still better than a plain typedef.

Yes it would. Please don’t do this.