I'm not trying to push Modula-3, just contribute facts.

> Hey I wouldn't mind having a Modula-3 compiler for my VAX, but I'm not
> going to be the one to port it.  Maybe we need a Modula-3 interpreter?  
> Or maybe we should instead port CVSup to perl so *everybody* can run it,
> including WinNT bozos.

I work with people who did a lot of the M3 work.  I asked one of 
them about porting.

There are two issues.  The first is the operating system dependent 
part.  That's usually the tough one, but we don't have to worry about 
that since Modula-3 is already running on NetBSD. 

The other area is the back end of the compiler.  It takes a new code 
generator for each CPU you want to run on.  I'm not a compiler wizard.  
I don't know how much effort that would take.  I think it should 
be not-to-hard if you aren't trying to squeeze the last drop of performance 
out of the target chip.  (The guy who wrote that code is good.  That 
area is probably pretty clean.) 

It might be possible to make a generic C backend.