Re: UNCOL = Uncool?

Peter Gammie <>
23 Oct 2000 21:41:11 -0400

          From comp.compilers

Related articles
UNCOL = Uncool? (SRS) (2000-10-19)
Re: UNCOL = Uncool? (Daniel C. Wang) (2000-10-22)
Re: UNCOL = Uncool? (2000-10-22)
Re: UNCOL = Uncool? (Peter Gammie) (2000-10-23)
Re: UNCOL = Uncool? (2000-10-23)
Re: UNCOL = Uncool? (Daniel C. Wang) (2000-10-23)
Re: UNCOL = Uncool? (Chris F Clark) (2000-10-23)
Re: UNCOL = Uncool? (2000-10-26)
Re: UNCOL = Uncool? (2000-10-26)
Re: UNCOL = Uncool? (Pred.) (2000-10-26)
[4 later articles]
| List of all articles for this month |

From: Peter Gammie <>
Newsgroups: comp.compilers
Date: 23 Oct 2000 21:41:11 -0400
Organization: University of New South Wales
References: 00-10-139 00-10-150
Keywords: UNCOL

Guy Steele floated the idea of SCHEME as an UNCOL in his Masters Thesis
of 1978 ("RABBIT: A Compiler for SCHEME" - MIT AI TR 474).

This work builds on his previous lambda-calculus stuff ("lambda: the
ultimate goto", &c.), and was the pioneer of the
compile-using-continuations stuff Appel et al pursued in the early 90's.

I'd suggest that, like the JVM, it's pointer representations are too
limited for it to be much of an (efficient) UNCOL. Unboxed types, for
example, would be impossible AFAICS.

Given how much success the Mercury and Haskell people have had, I concur
with you that something C-ish is where it's at.


"Daniel C. Wang" wrote:

> Also other than the original UNCOL paper (circa 1950 right?) the only
> other high profile UNCOL like language I remember was ANDF... I think
> ANDF failed for other reasons and that UNCOL's are not an obviously
> hopless idea...

Post a followup to this message

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