[gclist] Name that hypothesis

[Hans Boehm]

On Dec 3, 12:40pm, Nick Barnes wrote:
> Subject: [gclist] Name that hypothesis
> What is the name of the following hypothesis?
>
> 	"Most references in a system point backwards in time, i.e. from
> 	 younger objects to older objects."
>
> In the GC community we all know this, and many of us use it
> frequently, but we don't seem to have agreed on a name for it.
>
It seems to me that this is actually

1) Much stronger than required by a generational collector, and
2) False for a significant fraction of nonfunctional programs.

Counterexamples:

a) Programs that build lists or trees "front-to-back" or "top-down".

b) Programs that use mostly doubly linked structures.

I suspect neither is uncommon or necessarily undesirable.  (The fact that the
C++ standard library currently provides doubly- but not singly-linked lists is
likely to encourage the latter in C++.)  Also neither is necessarily going to
cause a generational collector to behave poorly, since there are still likely
to be few forward pointers between generations.

Hans



-- 
Hans-Juergen Boehm
boehm@mti.sgi.com