Is anybody using CPS compilation anymore?

In '93 I invested a lot of time reading papers and theses learning how
compilation by conversion to continuation-passing style (CPS) worked.
Later, though, I read Amyr Sabry and Matthias Felleisen's "Is
Continuation Passing Useful for Data-Flow Analysis?"
(http://www.cs.uoregon.edu/~sabry/papers/cps-flow-pldi.ps) which
proved, more or less, that using CPS doesn't really buy much of
anything, and that using it is more an issue of style than anything
else.


Coincidentally, reading PLDI, etc., papers over the last few years,
one sees fewer and fewer projects that are based on CPS conversion.
Is there any good reason for this? If one were to design a new
compiler, would using CPS conversion be a bad idea for any reason?
(That is, any reason other than the obvious objection that the CPS
intermediate form is harder to read than direct style formats.)


-Peter S. Housel- housel@ms7.hinet.net
--

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