3 ms·

I'm commenting primarily because I agree with fao_ that it's sad nobody's commenting on this post. Bottom line: the Lambda Papers (of which this is one) are a t

I'm commenting primarily because I agree with fao_ that it's sad nobody's commenting on this post. Bottom line: the Lambda Papers (of which this is one) are a true achievement in programming languages. Before these papers, there was a belief that lambda calculus was of theoretical interest only. Various interpreting automata (of which the SECD machine is probably the best known) were seen as too inefficient to be of practical use. Steele and Sussman demonstrated in the papers that procedure application subsumes many features of sequential programming languages, including assignment, GOTO, and object-oriented programming (the original plan was to build an actor language, until they realized that procedure application gives objects), and that you can build an efficient interpreter or compiler for the full lambda calculus. Anybody interested in programming languages can benefit from reading them.

fao_ 6y ago
Because the lambda papers were multiple submissions for a thesis (each one rejected(!)), sometimes comments slip in about certain things, and it's just generally fascinating to read.
- 082349872349872 6y ago
  p57 {LABELS} is interesting because historically it took a long time to get from LABEL to LABELS. One doesn't have to change much to add LABELS to a LISP 1.5, but because almost all the time LABEL suffices, I don't think anyone really felt the need. p63 {QUOTE shafts the compiler} is still a religious war to this day. p68 {WEBER} shows Steele & Sussman were writing pre-Moggi, but faced similar problems. (incidentally, the overruns in the box on p55 make me think that despite the right-justified text, the boxes were added after printing, via pen & ruler.)