Newsgroups: alt.lang.design,comp.lang.c,comp.lang.c++,comp.lang.lisp
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!gatech!newsxfer.itd.umich.edu!zip.eecs.umich.edu!caen!msuinfo!harbinger.cc.monash.edu.au!news.cs.su.oz.au!metro!news
From: maxtal@physics.su.OZ.AU (John Max Skaller)
Subject: Re: Reference Counting (was Searching Method for Incremental Garbage Collection)
Message-ID: <CzvtG1.818@ucc.su.OZ.AU>
Sender: news@ucc.su.OZ.AU
Nntp-Posting-Host: physics.su.oz.au
Organization: School of Physics, University of Sydney, Australia
References: <hbakerCzpJy7.ItH@netcom.com> <Czqu5L.6Cy@world.std.com>
Date: Sat, 26 Nov 1994 15:52:49 GMT
Lines: 38
Xref: glinda.oz.cs.cmu.edu comp.lang.c:117957 comp.lang.c++:100434 comp.lang.lisp:15819

In article <Czqu5L.6Cy@world.std.com> tob@world.std.com writes:
>hbaker@netcom.com (Henry G. Baker) writes:
>> Good ideas.  Now can you formalize these rules & prove conditions
>> under which any & all garbage will be collected?  I'm trying to
>> understand _provably safe styles_ of programming.
>
>Wow, we must really be approaching it from different perspectives.
>
>You seem to want a more formal(?) academic(?) statement, which I'm at a
>loss to frame. I probably have some implicit assumption, or assumptions,
>that you don't. To me, as a programmer, hearing or saying "there's a top
>structure that owns everything inside." expresses the idea precisely
>enough. I'm not sure what it is you want me to make more precise.

	Yep. Bertrand Russell had one of them too. Only he spent
ages trying to make it precise -- and still only found out just
in time to slip a page into his book saying "Woops -- the whole
thing is flawed!"

	You may think that a set is a collection of things,
but Russell showed vague, intuitive (naive) notions are misleading.

	See Ellis and Detlefs paper on GC in C++. They try to deduce what
subset of C++ will work with a conservative garbage collector.

	To compare a naive and formal statement, see the ARM
(naive) and the C++ Working Paper (attempts to be precise).
Examining those naive and hidden assumptions is important. 

	Helps in design too. Whenever I "feel" a piece of code
ought to work but I can't explain exactly why -- guess where
the bug pops up?

--
        JOHN (MAX) SKALLER,         INTERNET:maxtal@suphys.physics.su.oz.au
	Maxtal Pty Ltd,		    
        81A Glebe Point Rd, GLEBE   Mem: SA IT/9/22,SC22/WG21 
        NSW 2037, AUSTRALIA	    Phone: 61-2-566-2189
