@LeakyNun Nope. tio.run/nexus/jelly#@29oYHd4Qprx//8A

so [dyad]Ðf[nilad] is equivalent to [dyad][nilad]$Ðf, right? @Dennis

Yes. Ðf doesn't change the arity, so you get just another dyad and use it like usual.

But why does [dyad][monad]$Ðf work differently than [dyad]Ðf[monad]$?

Because [dyad][monad] and [dyad][monad]$ are not the same. The first applies the monad to the argument, the second to the return value.

I see

this reminds me: ÐṂ is almost useless because ÞḢ does the same thing (likewise ÐṀ)

That's not true. ÞḢ returns the keyed minimum, while ÐṂ returns the array of all keyed minimal elements.

ah right

in that case ÐṂ is even worse than I thought it was, as you need an Ḣ after it anyway most of the time ;-)

That said, all of these should be implemented in the same way Ðf is so they can take dyads as well.

(I'm mostly joking, sometimes you do want all the minimal elements)

Well, if you use ÐṂḢ instead of ÞḢ, that's on you. :P

yep

I find a really common mistake among Jelly programmers (including me sometimes)

is using a builtin that does what you want or almost what you want

when a solution which uses different bultins does things more directly

@Dennis I did use something like this in a proto-answer lol

(for example, using multiset intersection when a delete-not-in operation would do)

Yeah, with a large number of built-ins comes a vast of number of ways to pick the wrong ones.

I usually remember the stuff I implemented myself, but I routinely forget about atoms from pull requests.