🚀 go-pugleaf

In article <1ccaamINNgbt@uwm.edu> markh@csd4.csd.uwm.edu (Mark) writes:
>    I've been working on a new approach to functional programming that's
> resulted in quite a few interesting ideas.  The approach is based on the
> discovery of a fundamental correspondence between functional and  
imperative
> languages.  So I'd like to bandy about a few of the insights.
> 
> (1) Assignment-Let correspondence.
>    The basis of the correspondence is the following:
> 
>                 assignment statement <-> let statement.
> 
I like the idea of integrating functional and imperative programming, but  
don't think this particular correspondence holds up under close  
inspection. For instance, what do you do about procedures? E.g.

  let a = 0
  let inc cont = let a = a + 1 in cont
  in inc (inc (inc done))

clearly does not do three increments in the last line.
If your answer is: "expand the code of all procedures before
binding any variables" then you have essentially Forsythe.
If you treat instead the global state as an updateable function then you  
have essentially a denotational semantics of an imperative language.
In any case, something quite un-let-like goes on.

-- Martin Odersky