On 8/09/2015, at 4:57 pm, Manuel Gómez
We’re so used to understanding complex semantics and diverse, sometimes even obtuse styles of expression (pointfree, anyone?).
Correction. We are used to TRYING to understand ... Haskell long ago surpassed the opacity of much-maligned APL. It's *already* bl---y hard to read some Haskell code and removing signposts is *not* a good way to help the lost and bewildered.
Surely the mental effort of parsing a lambda or a do-block in a new, previously invalid syntax is a trivial matter for programmers accustomed to running type checking algorithms in our heads.
In the presence of type-level programming, my personal ability to run a time-checking algorithm in my head turns out to be embarrassingly limited. In any case, surely you have heard the proverb, "it's the last straw that breaks the camel's back"? As for me, I make enough mistakes that I cherish a syntax which rejects enough token sequences to catch many of my errors, while not rejecting so many that constructing a valid program is a puzzle. I'd be the last to say Haskell syntax is perfect, but it's d--ned good.