Monads in D

[Atila Neves via Digitalmars-d]

On Saturday, 11 June 2016 at 16:50:15 UTC, qznc wrote:
> On Saturday, 11 June 2016 at 14:26:20 UTC, Atila Neves wrote:
>> Why? Because I could, I don't plan on using this for anything 
>> serious. I think "with" is my favourite D feature right now. I 
>> also wrote the Writer and State monads (not that D needs them):
>
> I also tried this. Instead of Write and State, I tried to model 
> the Functor > Applicative > Monad > MonadFail type hierarchy. I 
> found no way for a good "template inheritance" construct. 
> Overall, I quickly forgot about it, because it looks ugly and 
> seems to have no advantage.

I was thinking of doing that as well and was going to model the 
hierarchy like so:

alias isFunctor = isInputRange;
enum isApplicative(alias T, U...) = isFunctor!T && is(typeof() { 
... }));
enum isMonad(alias T, U...) = isApplicative!(T, U) && 
is(typeof(() { ... }));


> Thinking more conceptually, Monads should be somewhat related 
> to input ranges, as both model a linear sequence.

Monads are related to input ranges because every monad is a 
functor, and input ranges are functors.

Atila