object expression and captured state in F#

Question

What makes the first implementation KO ?

type IToto  = 
    abstract Toto : unit -> unit

{ new IToto with  
      member this.Toto = 
             fun () -> () }

{ new IToto with  
        member this.Toto () = ()  }

Answer 1

In the compile representation, there is a difference between property of a function type, compiled as FSharpFunc<unit, unit> Toto { get; }, and a method taking unit and returning unit, compiled as unit Toto().

The first object expression implements a different interface:

type IToto  = 
    abstract Toto : (unit -> unit) // Note: Parentheses around the function type!

{ new IToto with  
      member this.Toto = 
             fun () -> () }