Type Inference behaving non intuitively with Functional Dependencies

[atravers]

@initial-mockingbird

[…] I was really wishing to have better type inference.

• Why does Haskell have functional dependencies?

  Because multi-parameter type classes can sometimes have ambiguous instances.

• Why does Haskell have multi-parameter type classes?

  Because the original (single parameter) type classes limited expressivity.

• Why does Haskell have type classes?

  To provide a means of reusing symbols with different types - overloading.

Having recently seen this:

• A functional programming language (suitable for education of functional programming) (2018) [from page 33 of 97]

and having already seen this some time ago:

• The implementation of practical functional programming languages (1991) [from page 63 of 144]

…is the very concept of a type class a mistake (or at least flawed)? Because it seems to me that the “Jenga-tower” of a type system Haskell now has is largely the consequence of that original decision.