Binary-coded decimals

I've been looking at these lately and can't figure out for the life of me why we would want to use them over a binary encoded number scheme.

I can understand that you can represent decimals to the exact location, but the overhead needed for this is just unacceptable.