Re: Tail recursion in object systems

fhd@Panix.Com (Frank Deutschmann) writes:
      I would be extremely interested to learn if it is possible to
      handle tail-recursion in constant space/resources in a message passing
      environment. ...


This is actually a bit humorous, since the ur-source of most of the past
two decade's work on tail-recursion is one and the same as the ur-source
of most work in true message passing, namely Hewitt's Actor model. See
Hewitt's paper "Viewing Control Structures as Patterns of Passing
Messages". This appeared in the Journal of Artificial Intelligence only
in 1977 (Volume 8, Number 3, pp. 323-364), but the content had been
floating around as MIT AI Lab Memos for a couple years before that. The
full tail-recursiveness of the Scheme programming language (Steele and
Sussman, 1975) was a direct response to Hewitt's work, attempting to show
how the same ideas could be embodied in a more conventional function-call
oriented linguistic setting rather than in the framework of message
passing. So, the question seems to have turned 180 degrees around in the
last 18 years, from "can we do this with function calling instead of
message passing" to "can we do this with message passing instead of
function calling".
--

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