I'm confused why it's said to be backwards (in the article and by you).
This paragraph shows that with a format string with an int format specifer, and then a string format specifier, the token list shows is exactly that, and the type it returns is a function that takes an Int and a String as input and returns a new string.
> For example, if the format string was "%d %s", the tokenization would be [IntFormat, NormalChar ' ', StrFormat], so the resulting type would be Int -> String -> String`. This might be confusing because it's a post-order traversal of the list, and might feel kind of backwards.
Isn't that exactly the order you'd expect?
I'm confused why it's called backwards, and why they say "post-order" traversal of a list, since I'm only used to hearing about traversing trees that way.
This paragraph shows that with a format string with an int format specifer, and then a string format specifier, the token list shows is exactly that, and the type it returns is a function that takes an Int and a String as input and returns a new string.
> For example, if the format string was "%d %s", the tokenization would be [IntFormat, NormalChar ' ', StrFormat], so the resulting type would be Int -> String -> String`. This might be confusing because it's a post-order traversal of the list, and might feel kind of backwards.
Isn't that exactly the order you'd expect?
I'm confused why it's called backwards, and why they say "post-order" traversal of a list, since I'm only used to hearing about traversing trees that way.