> > Can you please explain why 0 is a better "faked" CRC than 1? Both can be a
> > valid CRC so setting 0 is not a good way to detect it as disabled, right?
>
>I have no idea what detail you're talking about, but I recall that a CRC
>cannot become zero. (If this is a real CRC, not a checksum)
>
>Jörg

Just for interest, you may like to knopw about Adler32 "checksums". They
are mathematically as reliable as CRC32, but far simpler and faster to
generate. Designed by Mark Adler (of "gzip" fame) they come from a
reliable source.

I can see no reason why a CRC32 could not be 0. Unless there is some weird
reason in the spec explicitly prohibiting it removing it would just
increase the chance of collisions ...If you DO generate a 0, I wonder what
it should be replaced with? ...Just because I am interested in this kind of
thing.

BC

_______________________________________________
http://cool.haxx.se/mailman/listinfo/rockbox
Received on Sat Sep 3 22:35:28 2005