[MLton] Welcome to Kevin Redwine

[Stephen Weeks]

Welcome to Kevin Redwine (http://www.eecs.harvard.edu/~redwine/), who
has joined the MLton mailing list.  Kevin and Norman Ramsey have a
paper, "Widening Integer Arithmetic",

	http://www.eecs.harvard.edu/~nr/pubs/widen-abstract.html

that shows a nice way to implement small integers and words using
larger ones, i.e. exactly what we do in MLton when implementing, say
Int19 using Int32.

I asked them if they might like to try their stuff on MLton, and
explained a bit about our SSA IL, which is the right place to try out
their ideas.  Kevin indicated an interest, and here he is.

Welcome!


In an off-list email, Kevin said

> I'm not sure how much time I can spend on this at the moment, but I
> would be interested to get dynamic measurements for the naive vs DP
> algorithms using MLton.  Do you guys have much code that uses narrow
> integer arithmetic?

I don't know of any.  However, it occurs to me that it should be
pretty easy to turn many programs into test cases by putting
"structure Int = Int31 type int = Int.int" (or whatever size you want)
at the beginning of the program.  Provided the code doesn't overflow,
that should give a lot of data.  This might work especially well for
code coming from SML/NJ, which uses 31-bit integers by default.