Programming language projects

• Flame on object-oriented.
• Manifesto on JAR's next language.
• I follow the E programming language discussion, and contribute from time to time. One of my postings that might be of general interest is about Pascal, E, continuations, and monads.
• I was on a panel at MIT that discussed programming language design.
• Scheme + ADT's = secure OS. Idea described in "A security kernel based on the lambda-calculus," which was part of my PhD thesis.
• Mobot Scheme (at Cornell, with Bruce Donald). This is Scheme 48 running on a 4" by 6" embedded controller card (MC68000 with 1M RAM, .5M ROM). Byte code interpreter is in ROM. The board supervises a mobot with several other computers and a RWI B12 wheel base. Paper "Program mobile robots in Scheme" was in ICRA 92.
• Monad-parameterized Lisp interpreter. Monads (an invention of category theorists) give a good way of factoring the semantics of a language into essence and accident.
• Scheme 48, at MIT, Xerox PARC, and Cornell, with Richard Kelsey. See Lisp and Symbolic Computation 7(4):315--335, 1994.
• A reliable macro system for Scheme (at MIT; with Will Clinger; published in POPL 91) (originally conceived while working in Daniel Weise's group at Stanford)
• I was a teaching assistant for 6.821, MIT's graduate programming language core course; I wrote software and course notes, and generally helped develop the course (with Dave Gifford). The course has evolved a bit since I was last involved, as has my thinking on programming language pedagogy.
• Consultant to VAX Lisp product development group (DEC's Common Lisp)
• Editing the thrice-revised Scheme Report (MIT; with Will Clinger)
• Pseudoscheme (at MIT) -- Scheme embedded in Common Lisp
• T3 project (at DEC WRL, with a cast of thousands) -- hairy native code Scheme compiler. Paper in SIGPLAN Compiler Construction 1986.
• T project (at Yale, with Norman Adams, Kent Pitman, and Jim Philbin) -- used Steele's S-1 Lisp compiler and ideas from NIL to build the first compiler-based Scheme implementation. Paper in 1982 ACM Lisp Conference (which wasn't called that).
• NIL (New Implementation of Lisp) (at MIT), where JonL White taught me all I know about repulsively clever Lisp implementation techniques.
• Snostorm (SNOBOL4 preprocessor, at Yale). SNOBOL, with a modest set of improvements, and Algol 68 were my favorite languages before Drew McDermott showed me their similar but cleaner cousins Lisp and Scheme.
• SNOBOL4 port to Sigma-7 (at Grand Valley State Colleges, Michigan)

Editorial

Even after 50 years of hard work by many smart people, programming is still much more difficult than by all rights it should be.

I attribute this in part to the failure so far of creative individuals to find good methods and solutions, and in part to the failure of the software industry (and one company in particular) to be aware of and exploit known good practices.