Re: OCCAM, JOYCE, SUPERPASCAL AND JAVA

[Oyvind Teig <Oyvind.Teig@xxxxxxxxxxxx>]

1. I like the idea to have memory allocated as it is now, 
   and expand with a REC word. I have always bragged about 
   occam knowing exactly how much workspace it needs!

2. I don't know whether Roger Henriksson's "Scheduling 
   Garbage Collection in Embedded Systems" 
     http://www.cs.lth.se/~roger/thesis.html
   contains any valuable info. Probably not, but this 
   is a paper that Bill Foote recommends for reading, in
     http://www.nist.gov/itl/div896/emaildir/rt-j/msg00483.html

3. Another thing, Ken Thomson says in an interview in IEEE Computer
   May 1999, p62 that Limbo "doesn't have big mark-and-sweep type 
   garbage collection. It has reference counting. If you open 
   something and then return, it's gone by reference count. Then you
   don't have high and low watermarks because 99 percent of the 
   garbage goes away as soon as it is dereferenced. If you store a 
   null in a pointer, it chases the pointer and all that stuff 
   goes."
   ..
   "So, again, it's pragmatic. It's not the theoretical top-of-the-
   line garbage collection paper. It's just a way of doing it that
   seems to be very, very effective."

I guess these points are almost orthogonal to your problem,
Peter, but then it seems like recursion and dynamic memory 
handling converge. Do they, or have I misunderstood?
-- 

|=======================================|==========================|
| Oyvind Teig                           | oyvind.teig@xxxxxxxxxxxx |
| Navia Maritime AS, division Autronica | oyvind.teig@xxxxxxxxxxxx |
| 7005 Trondhem     |    http://www.navia.no | Tel:+47 73 58 12 68 |
| Norway            |http://www.autronica.no | Fax:+47 73 91 93 20 |
|===================|========================|=====================|