Re: Are there any studies or reports about the analysis

> A six month old message... :-)
> 
> Chin-Yang Lin wrote:
> > I am studying the traditional gc algorithm - reference counting(RC). A major weakness for RC is the inability to reclaim cyclic garbage. Many studies proposed their cyclic reference counting algorithms to deal with this problem in different efficient manners. 
> 
> I was wondering if somebody could provide some (online, please :)
> references about collecting cyclic structures in a Reference Counting
> framework.
> 
> I'm working in an object database backend. I'm currently doing GC using
> reference counting, and I'm leaking objects in cyclic structures. This
> is a rare situation, but the uncollectable garbage grows with time.
> 
> I would like to be able to collect that garbage without examining a
> multimillion object database :-), object by object (let's say,
> mark&sweep is not an option)
> 
> I could change GC algoritm easily, in my project. So, I'm not committed
> to Reference Counting. But I need an algoritmo which doesn't need to
> revise all objects in database, even in a "generational" way. My
> advantage is that my mutator and the GC system can fully cooperate,
> 100%. That situation should be very helpful.

I know of no comprehensive on-line reference but there's a pretty
comprehensive (IMHO :-) section in my book, Garbage Collection, Wiley 2000.

Richard Jones