Path: utzoo!attcan!uunet!ginosko!gem.mps.ohio-state.edu!wuarchive!mailrus!cwjcc!gatech!hubcap!billwolf%hazel.cs.clemson.edu
From: billwolf%hazel.cs.clemson.edu@hubcap.clemson.edu (William Thomas Wolfe, 2847 )
Newsgroups: comp.sw.components
Subject: Re: Real-time Garbage Collection
Message-ID: <6559@hubcap.clemson.edu>
Date: 24 Sep 89 00:36:03 GMT
References: <194000001@hollin>
Sender: news@hubcap.clemson.edu
Reply-To: billwolf%hazel.cs.clemson.edu@hubcap.clemson.edu
Lines: 23

From ds@hollin.prime.com:
> Sometimes, [...] dynamic reclamation (garbage collection) is 
> truly necessary (to avoid unnaturalness or inefficiency).  A
> useful parallel is the need for both stack and heap data storage in
> general-purpose programming languages.

    I think we've established that managed and unmanaged storage
    paradigms can coexist, and that components which manage their
    own storage can avoid the inefficiencies of garbage collection.

    We also know that the user MUST be involved in storage management,
    if for no other reason than to decide which data structures to
    throw away in the event of a storage crisis.  

    What I'd like to know is: Is there a hard counterexample which 
    will demonstrate that situations exist in which the inefficiencies
    of garbage collection cannot be avoided?  Someone has already outlined
    the properties that such a counterexample should have, and mentioned
    a "gut feeling" that such a counterexample exists.  Unfortunately,
    I have a similar "gut feeling" that GC is never worth the cost. 


    Bill Wolfe, wtwolfe@hubcap.clemson.edu