Re: call by name, was silly question: prefix vs postfix ops

our moderator wrote:
>[ [... I]t does seem odd that as best I can tell, nobody said
>anything about the complexity of implementing call by name until
>Jensen's paper. -John]


Worth remembering that for most people in the early '60s Algol
60 was not a serious contender as a practical programming language,
and so the complexity of implementation was barely an issue. Algol
was primarily intended as a language in which to communicate
algorithms, so that we could read code in CommACM or CompJ and
transcribe it into the autocodes available on our computers. Those
who were trying to describe the latest wrinkle in matrix inversion, or
quicksort, were mostly quite horrified when "subversive" programmers
started to explore the corners of Algol.


Later, of course, some of the same mistakes were repeated with
Algol 68. Care was taken that algorithms could be described cleanly,
clearly and efficiently, but difficulty of implementation was only a
minor issue, until the 68R dialect showed what could be done, and
rather more interest [though still perhaps not enough] in
implementation was taken with the revision. The assumption seems to
have been that compilation techniques would eventually catch up; which
I suppose they did.


--
Andy Walker
Nottingham

Return to the comp.compilers page.
Search the comp.compilers archives again.