home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!cs.utexas.edu!sun-barr!olivea!spool.mu.edu!darwin.sura.net!aplcen.apl.jhu.edu!aplcenmp!hall
- From: hall@aplcenmp.apl.jhu.edu (Marty Hall)
- Newsgroups: comp.lang.lisp
- Subject: Re: Reference counting (was Re: Yet another Lisp Interpreter Available (RefLisp) Long)
- Message-ID: <BzMD1E.EGp@aplcenmp.apl.jhu.edu>
- Date: 21 Dec 92 17:10:25 GMT
- References: <1gk60iINNjg8@iraul1.ira.uka.de> <1992Dec17.200901.20942@uk03.bull.co.uk> <1gr5ejINN3k5@early-bird.think.com>
- Organization: AAI Corp AI Lab, JHU P/T CS Faculty
- Lines: 19
-
- barmar@think.com (Barry Margolin) writes:
- >bbirch@hemel.bull.co.uk (Bill Birch) writes:
- >> - RefLisp has reference counting garbage collection, that means
- >> memory is reclaimed as soon as it is un-used. It is a "non-stop"
- >> system, in that there is never any halt for garbage collection (GC).
- >
- >This is a frequent claim that is simply not true. A reference counting
- >system simply changes the time and distribution of the GC delays, but it
- >doesn't eliminate it. For instance [...]
-
- A PostScript version of a good (in my non-expert opinion) survey of GC
- techniques is available by anonymous FTP from cs.utexas.edu, in
- /pub/garbage/gcsurvey.ps.
-
- Wilson, Paul R., "Uniprocessor Garbage Collection Techniques", to appear
- in _Proceedings of the 1992 International Workshop on Memory Management_.
-
- - Marty
- (proclaim '(inline skates))
-