Interesting/Related Projects

New additions to this list

• ABCL
• MOVITZ
• POPLOG / POP-11

Virtual Machines, etc.

"Fundamentals of New Computing" (FoNC) -- see:

• http://piumarta.com/pepsi/- Ian Piumarta's experimental framework for dynamic languages.
• http://vpri.org/mailman/listinfo/fonc
• http://vpri.org

A possible starting point for TUNES?

Parrot (2002-present) is a lower-level VM designed mainly for Perl and other dynamic languages (Ruby, Python, Lisp/Scheme, Smalltalk, etc.).

• Not our ideal framework because it's not helpful for metaprogramming, however...
• It has the potential to consolidate the infrastructure of several popular languages in a more efficient implementation that offers "real" garbage collection, thus providing short-term practical benefits.
• Should allow mixing code written in different languages... including common-language libraries... which should help to consolidate some of these languages eventually.

LLVM - Low Level Virtual Machine

Programming Language projects

PyPy - Python written in Python (2003-present)

A few notes:

• JIT compilation (as in Psyco - see http://psyco.sourceforge.net/doc.html under "How does it work?")
• Support for lazy evaluation, logic programming, persistent objects
• C, LLVM, and .NET CLI back-ends
• Javascript & Prolog front-ends begun...

Ruby 2.0 - several implementation efforts underway:

• jRuby (Sun Microsystems) - JVM-based, of course. Version 1.0, released June 2007, supports Ruby 1.8.5.
• Rubinius (Evan Phoenix)
• MRI (Matz's Reference Implementation)
• Microsoft Ruby (John Lam)
• Cardinal (Kevin Tew) - uses the Parrot VM

Pat Collison's CROMA Lisp - Said to be a cleaner, re-invented Lisp dialect. Won him two awards and no doubt helped him get into MIT.

If it's ever released, we'd like to play with it.

Squeak Smalltalk explores a lot of ideas related to Tunes.

Clean Slate Smalltalk (2003-2005)

Brian Rice says the project is dead, mainly due to poor performance.

Max (2002-2004, 2007) - a reflection framework for Common Lisp. It's clear the authors put a lot of thought into this, but it hasn't generate the interest they hoped for. The easiest way to find out what it's all about is to skim the Documentation.

As of August 2007, Tril has resumed work on Max.

Retro (1998-2002) showed that driver writing and other "OS stuff" is relatively easy, but FORTH is inadequate for our purposes.

After 2002, work continued on RetroForth and its descendants. Perhaps most notable is Factor, which has more in common with Lisp than with Forth.