[MLton] ML Workshop, sML Evolution, and CUFP trip report

[Andreas Rossberg]

skaller wrote:
> 
> Gak, I don't understand any of that. I'm implementing typeclasses
> in Felix at the moment:
> 
> (a) AFAICS there is no record to pass around: it's a whole program
> analyser like MLton, the instance functions are known at compile
> time and called directly (with no overhead).

You are talking optimizations, while the argument was about semantics. 
It's like saying that functors aren't functions (over modules), just 
because they can be applied at compile time in certain compilation 
approaches.

> (b) there's no type inference: Felix does already
> support overloading, so this may be masking the kind of extension
> ML would require

Without some inference, type classes would just be a restricted form of 
(first-class?) modules. I have no idea what use that would serve. Note 
that overloading resolution already is a form of inference.

- Andreas