On Fri, Jun 22, 2001 at 04:04:27PM -0400, Dave Sill wrote:
> "Dave Sill" <[EMAIL PROTECTED]> wrote:
>
> >If the input numbers are not fairly random, then a modulo hash is not
> >a choice.
>
> Not a *good* choice.
>
> >>Unix file system inode numbers are not truly random. Therefore, it's
> >>wise to choose a prime conf-split.
>
> BTW, I modified my modhash program to read numbers from stdin, fed it
> lists of real, live inode numbers, and guess what? It still makes no
> difference whether you use a prime hash or not.
You'll need local data to write the optimal hash function. The prime
doesn't guarantee anything but it's usually a good compromise for the
general case (see my earlier reply to Krieger).
Jörgen
