It uses explicit regions instead of garbage collection.
Explicit regions are both very simple and very fast, but how far can one get with them? I want to find out, so I am developing this interpreter.
So we've come full circle.
All research of Interlisp-D, Smalltalk and Mesa/Cedar at Xerox PARC:
https://archive.org/details/bitsavers_xerox
Stéphane Ducasse also maintains the original Smalltalk manuals on his legacy books:
http://stephane.ducasse.free.fr/FreeBooks.html
Oberon at ETHZ:
http://www.ethoberon.ethz.ch/books.html
If you want to see how the latest incarnation of Oberon used to look like:
http://www.progtools.org/article.php?name=oberon§ion=com...
For Modula-3 it is a bit harder, because of DEC/Olivetti being acquired by Compaq, which was then acquired by HP. So lots of sites are full with broken links.
So the older books are a better source than the Internet.
But still you can get some info here:
http://ftp.labs.hp.com/ftp/pub/dec/SRC/
http://www-spin.cs.washington.edu/
Burroughs overview:
http://www.smecc.org/The%20Architecture%20%20of%20the%20Burr...
It lives on as Unisys MCP mainframes. There are some more informations when searching for it.
Algol-68RS used at Royal Navy:
https://en.wikipedia.org/wiki/ALGOL_68RS
This is just a small taste. If one cares about history of computing there are lots of other resources scattered around the web and libraries.