On 2 Feb 2002, Cees de Groot wrote:

> Scott A Crosby <crosby at qwes.math.cmu.edu> said:
> >So, do keep these in mind, if you're thinking of translating 'Morph' into
> >several languages.
> >

> With all respect, but form follows function here. If (a big if) it
> would be deemed necessary to have variants of selectors in all
> languages and that would break some limits, well then we need to kiss
> goodbye to the limits...

Even with an image rebuild that fixes the flaw, going over 2000 would be a
bad idea. Fixing the flaw avoids the brick wall at 4000, but it can't
remove the performance degradation once you get past 2000.

The only way to fix this is to have more random hash bits available.

>
> (how many distinct selectors are there at present anyway?)

25k distinct selectors implementing 40k methods.

Scott