[MLton] straw-man packaging proposal

[Stephen Weeks]

> I don't want to rain on anyone's parade here, but I really do
> believe this is one wheel which should be re-invented only with
> great trepidation.

I completely agree, at least in the sense in which my post seems to
have been interpreted.  I also largely agree with Wesley's response
that he sent a couple of months back.  Packaging was clearly a bad
name for what I intended, because both responses interpreted my mail
very differently than it was meant.

What I didn't make clear in my mail, and plan to elaborate on in a
future mail when I have time, is that I do not want replace host
packaging systems (e.g. the Debian packaging system).  Rather, I want
to provide mechanisms that make sense within the context of SML+MLBs
for combining SML libraries together to form whole programs.  MLBs are
useful, but I think they are not a complete answer.  To sum up my idea
in a single phrase, I would like to view SML libraries (defined by
MLBs) as functors on bases, and to view assembly of a program as
linking together those functors.  MLBs are not capable of expressing
such linking, so I believe we should either extend MLBs or design a
linking language on top of them.

Whatever mechanism we design will live within the host packaging
system that handles all of the grungy details that Wesley and Cynbe
mentioned.