[gclist] Garbage collection FAQ.

[Stefan Monnier]

> Is it a good idea to valorize reference counting by putting it first and
> at the same level as the other approaches?  The myth to the effect that
> "GC is reference counting, reference counting sucks, therefore GC is
> bogus" is pretty widespread.

As far as I know GC is more understood as "it's slow and freezes my emacs
every so often". That's why many people "like" reference counting (it's
probably the most widely used GC algorithm, especially for non GCed languages
like C++).
Furthermore, it is an alternative technique and is worth mentionning just
like mark&sweep and stop&copy. Reference counting is not perfect, far from
it, but it can be better suited than tracing-GC in some cases.

> What is the disagreement?  Other than the pig-in-the-poke problem, and
> all other things being equal, copying has been a win over mark-and-sweep
> as far as I have seen in practice.

Some people are convinced of the opposite. And some people don't really
care. It's all a matter of tradeoffs and there is no "one size fits all".


	Stefan "who admits being more of a copy-bigot than a mark-bigot"