AccessFunction is only a syntactic sugar

[tunguski]

AccessFunction is only a syntactic sugar

[jfmengels]

I'm not sure what you are suggesting.
Are you suggesting that

fn = .field
{- should generate
  Lambda [ Variable [ "x" ] ] (Access (Variable [ "x" ]) [ "field" ])
-}

or that

fn = \x -> x.field
{- should generate
  AccessFunction [ "field" ]
-}

?

Either way, my opinion is that things are fine the way they are now. Having the distinction allows for better analysis. For instance, in a tool I'm working on that analyses your code and reports improvable patterns, I report this exact behavior that could be simplified, so that the user can fix it (see SimplifyPropertyAccess). If the AST does that simplification for me, I have no way of knowing how it is written and reporting it to the user.

[tunguski]

Hi @jfmengels, you are totally right. Described proposal was based on a specific use case, and as you pointed out it is not proper for general parsing library.

Thanks!

Ps.: it was about removing AccessFunction and to directly write it using simplified type.