From: "M. Edward (Ed) Borasky" <znmeb@...>
Date: 2007-12-27T05:28:02+09:00
Subject: Re: Destroying an Object

tekwiz wrote:
> On Dec 26, 11:18 am, Tim Hunter <TimHun...@nc.rr.com> wrote:
>> Ken A. wrote:
>>> Tim Hunter wrote:
>>>> Ken Awamura wrote:
>>> Does ruby's garbage collector really work, or it's as lame as .NET
>>> garbage collector?
>> It really works.
> 
> That's not entirely correct.  The newest .NET GC uses an algorithm
> that is actually more efficient than the inefficient Mark and Sweep GC
> algorithm that Ruby uses.  However, Ruby's GC is much less buggy
> than .NET's in my experience.
> 
> --
>      Travis
> 
> 

Well ... I think some universities still give degrees for garbage
collector research, although hopefully not PhDs any more. ;)

But seriously, folks, there is so much research out there that
statements like "as lame as .NET" or "the inefficient Mark and Sweep GC"
aren't particularly informative. In any event, there are efficient Mark
Sweep GCs, inefficient Stop Copy GCs, generational GCs, compacting vs.
non-compacting GCs, GCs that attempt to work with the OS memory
management rather than independent of it, etc.

For Ruby, I suspect the jRuby/JVM garbage collector is probably the best
one out there, simply because it has more person-decades invested in it
than the ones in MRI or KRI. So there are indeed "opportunities" for the
GC experts.