Martin McClure wrote:

> One more possible change: It was brought up a few months back that
> the number of object header bits devoted to identity hash (10 or 11
> bits, IIRC) wasn't enough to allow efficient hashed identity
> collections that were very large. It would be nice to have more bits
> for this purpose. On the other hand, I don't really want to increase
> the footprint for small-memory systems, and I don't want to give up
> image portability either, so I'm not sure I really want to propose
> making the header larger.
Improving the hash would be well worthwhile, especially if it can be
done for little cost. I don't think that a change that requires adding
more header size would be sensible.

Changing the hash prim to multiply the value from the header is
certainly plausible, as would be making it look up a 'better' value via
a table (4kbyers would cover all 1k curent hash indices; would a table
of 1k prime numbers be better? Or 1k spread out not-quite simple multiples?)

As you pointed out, strictly speaking this does not require an image
format change, but I guess it conveniently fits in the general area of work.

tim