Is there any good reason for the standard prelude defining
infixr 0 $, $!, `seq`
ie, making $ and $! right associative? I understand that
for `seq`, we like
x `seq` y `seq` z = x `seq` (y `seq` z)
but why
f $ x $ y = f $ (x $ y)
I'd rather prefer
f $ x $ y = (f $ x) $ y
ie, ($) is like application by juxtaposition, but changes
the precedence level (not the associativity). Whereas `f $
x $ y' may be rare, I sometimes like to write `f $! x $! y'
meaning `f x y' with strict application.
Cheers,
Manuel
PS: If this has been discussed before, could someone please
give me a pointer to the discussion?
