[MLton] Re: MLton and the Computer Language Benchmarks Game (the Shootout)

[Matthew Fluet]

On Fri, Sep 16, 2011 at 3:26 PM, Magnus Rattfeldt
<magnus.rattfeldt at gmail.com> wrote:
> I'm curious why MLton (and SML in general for that matter) is missing
> (was removed) from the Computer Language Benchmarks Game
> (http://shootout.alioth.debian.org)?

You would have to ask the maintainers of the Benchmarks Game for the
definitive answer, but I suspect that a lack of implemented benchmarks
prompted the removal.  In a previous iteration of the Game (under the
Shootout moniker), SML/MLton did have implementations for most, if not
all, of the benchmarks.  But, they have (slowly) cycled old benchmark
programs out and new benchmarks in, and, presumably, after a
language/implementation falls below some threshold of implemented
benchmarks, it is removed.

As for why MLton/SML doesn't have implementations for the new
benchmarks, it is the standard lack of time/motivation.  I,
personally, find the Benchmarks Game pretty poor for evaluating MLton.
 Microbenchmarks, like these, are helpful for compiler writers (they
identify gross inefficiencies), but they miss the point for real users
(who don't write microprograms) and miss the point for whole-program
optimizations.  Every compiler is "whole-program" if the program is 15
lines long.