Re: Fun-O Basic Edition Compiler

[Bruce Hoult <address@hidden>]

In article <3C2A9248.8CB924BF@quiotix.com>, Jeffrey Siegal 
<jbs@quiotix.com> wrote:

> Bruce Hoult wrote:
> > What I'd expect.  Stalin is known to be good.
> 
> An important point is that there are no declarations at all.  None are
> needed.  This, of course, is a trivially small program, making easy work
> for global type inference (though the other languages didn't seem to
> have such an easy time with it).

d2c often manages to do without declarations, including on simple 
recursive functions.  Somehow this 2-level recursion throws it though.

Most of the rest of the time declaring nothing more than function 
argument and return types is enough, and that's a) necessary anyway for 
dispatch in Generic Functions, and b) good practice for documentation 
purposes anyway.


> However, with some minor tuning effort, I get similar efficiency
> out of Stalin on larger Scheme programs as well (50-100KLOC)

How does it do on optimizing object systems built from raw closures (or 
otherwise...)?

-- Bruce