Path: utzoo!attcan!utgpu!news-server.csri.toronto.edu!rutgers!cs.utexas.edu!rice!news
From: wright@utopia.rice.edu (Andrew Wright)
Newsgroups: comp.lang.lisp
Subject: Re: Weak Reference in Common Lisp
Message-ID: <1990Aug18.190312.25603@rice.edu>
Date: 18 Aug 90 19:03:12 GMT
References: <81084@aerospace.AERO.ORG> <5819@hplabsz.HPL.HP.COM> <44659@cornell.UUCP>
Sender: news@rice.edu (News)
Distribution: usa
Organization: Rice University, Houston
Lines: 24

In article <44659@cornell.UUCP> murthy@algron.cs.cornell.edu (Chet Murthy) writes:
>>In article <81084@aerospace.AERO.ORG> srt@aero.org (Scott "TCB" Turner) writes:
>>>Is there a mechanism in Common Lisp (specifically AKCL) for weak
>>>reference?  By weak reference I mean a reference that won't prevent
>>>garbage collection if no other references exist.

>I know that there is such a facility (weak pointers) in the SML-NJ
>(Standard ML of New Jersey) system.

I think Chet must be referring to "weak types" of standard ML.
SML's weak types have nothing to do with what the original poster intended.

SML may be thot of, loosely, as lisp with a static polymorphic type system.
Reference cells are given a "weak" or "imperative" type in order that
they interact properly with polymorphism.  The semantics of imperative types
are presented in "The Definition of Standard ML", by Milner, Tofte and Harper,
MIT Press 1990, and are explained in more detail in Tofte's Phd thesis
(Edinburgh).  The term "weak" refers to an extension of "imperative" as
implemented by Appel and MacQueen in Standard ML of New Jersey, and
described only very briefly in the release documentation.


Andrew K. Wright        Computer Science, Rice University
wright@rice.edu         Houston Texas  77251-1892