[MLton] MLton library project

[Stephen Weeks]

> > Makes sense.  We definitely did not intend for others to assign
> > copyright.  I've tweaked doc/license/MLton-LICENSE (and hence
> > http://mlton.org/License).
> 
> "the above copyright notice" doesn't make much sense. 

True.  To be clear, what is actually being shown on
http://mlton.org/License is MLton-LICENSE-only, which is MLton-LICENSE
with the copyright notice removed.

> It's also not clear if you need to keep copyright notices in
> compiled binaries.

I'm afraid I didn't write the license and can't change it, at least
not without the original copyright holder's (NEC's) permission, which
is not really obtainable at this point.  So, that ambiguity will
remain.

> You reallly should give a precise copyright statement in the
> MLton-LICENSE file that fulfills the requirements of the license.
> Or you should change the license so that it doesn't require that
> downstream developers compile such a list of copyright notices
> on their own.

The MLton-LICENSE file currently include our usual copyright.

  Copyright (C) 1999-2005 Henry Cejtin, Matthew Fluet, Suresh
     Jagannathan, and Stephen Weeks.
  Copyright (C) 1997-2000 by the NEC Research Institute

I see two ways forward.

 1. Require that all contributors who want to assert copyright modify
    MLton-LICENSE, extending the copyright notice.  Go back to the old
    http://mlton.org/License, and add a note explaining that there is
    no intention to assigning copyright and that contributors can add
    to the notice. 
 2. Require everyone to contribute under a slightly different license,
    not the MLton-LICENSE, but that is compatible with the
    MLton-LICENSE.

(1) seems better to me.  What do people think?  In particular, Ville,
is (1) OK with SSH?