The FTP fiasco

Isn't it hilarious? But the real question is, why are Haskallers so stupid? seriously, it's almost like they want to shit all over what took them decades to build...

fuck the police

>>13,15
Foldable is just ToList, Traversable is just ToListApplicative. For a language that is so crazy about linked lists, the adoption of those classes into Prelude is only logical. It would make sense for Lisp, too, but Lisp just cannot into typeclasses.

>>42
It's also an examplar abstraction creep.
See, a class for fold sure does help a lot: pretty much every agelbraic type can be folded and you can even mechamically derive fold for them too, but why the hell does Foldable have fucking length in it?

>>43
fold has a very different type for each algebraic type though.. it's not something you can abstract over with a class

>>43
because length = foldl (const . (+1)) 0

>>45
Well I mean why just let List keep length and call the Foldable equivalent total or something? Because the name length implies some sort of sequential linear structure, while a Tree can be Foldable too.
>>44

it's not something you can abstract over with a class

You can if you restricted it to type constructors with kind *->*, and Foldable does just this.
class Foldable f where foldr::(a->b->b)->b->f a->b

Doesn't stop Foldable from being a convoluted OVER-ENGINEERED piece of shitware.

>>43
Because Foldable is ToList. If a data structure is coercible to a list, then there obviously is a way to define length for it.

>>46
Being a member of Foldable already implies some sort of sequential linear structure, whether it's a tree or not is irrelevant. If it's listable, there is length of it.

>>48
No it doesn't you monkey, it just says there is a structure-preserving map from some initial algebra to some other algebra. It implies nothing of there being a measure that corresponds to the length.

data Expression = Number Int
| Add Expression Expression
| Minus Expression
| Mult Expression Expression
| Divide Expression Expression

>>46
that's not a proper fold though, as 47 its a shittly listlike structure. Don't let them use 'advanced' names like Fold trick to you into thinking there's anything clever going on here. It's java level unprincipled bad design.

>>49
don't conflate category theory ideas like inital alegebras and folds with the cappy 'likelist' typeclass they've come up with and happened to call fold.

exercise:
(1) Define a foldable instance on your Expression a type (change it so Number Int is Number a).
(2) Remember how in category theory the identity function is an example of a fold? Define that using foldr :: (a->b->b)->b->f a->b

1 is possible of course but 2 is not, because it's not a real fold.

fuck what these people are doing to destroy this once beautiful language. Seriously? Am I the only one seeing this happening? They keep moving further and further away from mathematically justified things.

>>49
You're converting a datastructure to a finite list. All finite lists have lengths. Which fucking part don't you understand, caveman?

>>51

mathematically justified things.

Yeah, like Num. All of my category theory books say that numbers must have functions for negation and conversion to strings.

>>53
I agree Num is and has always been shit. I don't think that invalidates my other point though.

real fold

you mena CATAMORPHISM

Haskell is deprecated.
Install Rust today!

fold my anus

CATAMORPH MY ANUSMORPHISM

>>58
I prefer Ana^_lmorphisms.

it's just a trick to invalidate the old books and sell new books