[Haskell-cafe] Nested (=>) (Was: On finding the right exposition...)

[Ben Franksen]

Am 04.10.21 um 18:12 schrieb Henning Thielemann:
> On Mon, 4 Oct 2021, Ben Franksen wrote:
>> Talking of inconsistencies, Haskell is not without some of those. For 
>> instance, I never understood why (->) is right associative in types, 
>> but (=>) is not and you are instead supposed to pack constraints in a 
>> tuple. (It works in some simple cases but not consistently.)
> 
> I use nested (=>) frequently. GHC allows it in type signatures but not 
> in super-class constraints. Would be cool to be allowed everywhere, as 
> it supports the "terminator syntax style".

I am having second thoughts. The analogy of => with -> breaks down when 
you consider that the order of constraints never matters. Despite the 
tuple-like notation, I guess the better intuition is that constraints 
are always sets, "," is union, and simple constraints are singleton 
sets. This also works better when using constraint synonyms such as

   type OrderedNum a = (Num a, Ordered a)

I can write

   f :: (OrderedNum a, Floating a) => a -> a

which would otherwise be a nested tuple?

Cheers
Ben
-- 
I would rather have questions that cannot be answered, than answers that
cannot be questioned.  -- Richard Feynman