Why foldRight is beautiful Jun Tomioka (M3, inc.)

M3, Inc. Jun Tomioka Twitter: @jooohn1234 Github: jooohn Love: Our very first baby Had great paternity leave!

No content

foldLeft / foldRight

trait List[+A]

foldLeft[B](z: B)(op: (B, A) => B): B A A A A A B A A A A B B OP ...

foldRight[B](z: B)(op: (A, B) => B): B A A A A A B A A A A B B OP ...

So what’s the difference between them?

Suppose you define your own Linked List

foldLeft would be like this

foldRight would be like this

This naive foldRight can cause stack overflow

If so, why can foldRight be beautiful?

Let’s implement “map” with foldLeft

A A A A A List[B] A A A A OP ... List[B] List[B]

Let’s implement “map” with foldLeft

foldRight

Let’s implement “map” with foldRight

A A A A A List[B] A A A A List[B] List[B] OP ...

Why is it that natural to write “map” with foldRight?

Immutable recursive data structure

Bigger part holds smaller part

We must create them from smaller part to bigger part

This is the folding order of foldRight

foldRight folds items by its creation order

Why foldRight is beautiful ● Immutable recursive data structure is built from smaller part to bigger part ● In this sense, foldRight folds items from smaller part to bigger part rather than from right to left ○ Is quite natural to treat immutable recursive data structure

Thanks!