Presumptuous monadism

[stuartpb]

I passed this input to pointfree.io:

latterLetter1d alphabet letter = (elemIndex letter alphabet) `further` latter alphabet

This was the output:

latterLetter1d = ap (flip . (further .) . flip elemIndex) latter

Is there no pointfree way to express this without importing Control.Monad? Or, if monads are really the cleanest way to do this, is there a formulation of it without so many dots and flips (#23)?

[josephcsible]

Is there no pointfree way to express this without importing Control.Monad?

Yes, ap is the same as <*> in this case, and the latter is in the Prelude:

latterLetter1d = flip . (further .) . flip elemIndex <*> latter

Or, if monads are really the cleanest way to do this, is there a formulation of it without so many dots and flips (#23)?

Yes:

latterLetter1d = flip (flip (liftM2 further) latter . elemIndex)

[bmillwood]

Nowadays I think you don't need to import Control.Monad to get these instances, so I don't think it's bad to use the monad interface here.

[josephcsible]

No, you still do need to import Control.Monad, even in GHC 8.8.1.

[josephcsible]

Hello?

[bmillwood]

oh, I misunderstood this at the time, sorry (just happened to revisit it now). The issue is not the instance itself, but the function ap.

Nevertheless, I don't personally feel excited to work on this. It seems like using standard base imports is often going to result in better code, so at most I'd want this to be an option, --prelude-only or something.

[josephcsible]

It seems like using standard base imports is often going to result in better code

For other things like join and liftA2 I agree, but for ap in particular, is there ever a reason to use it rather than <*>?