rsync and md4 (my final comments)

[Perry E. Metzger]

"David F. Ogren" writes:
> 1. I think that you will agree that MD4 will work fine for Mr. Tridgell's 
> program, irregardless of your criticisms.  He specifically stated that he 
> was not concerned about intentional collisions, only random ones.

If one is concerned about speed and doesn't need a cryptographic
checksum, a long CRC will be far, far faster and will do fine.

As soon as one starts talking about using cryptographic checksums,
there is no point in using them unless one really wants the
cryptographic protection.

Perry