Re: Unique number generation

paul_at_atom.sbrk.co.uk
Date: 04/29/05 

Date: Fri, 29 Apr 2005 14:41:18 GMT

In article <d4tf8j$8si$1@nntp.itservices.ubc.ca>, Unruh wrote:
> paul@atom.sbrk.co.uk writes:
>
>>I'd like to be able to generate some secure unique numbers. A simple
>>implementation might be val = encrypt(secret, seq_no), however that
>>doesn't ensure that the numbers will be unique. Ok, I could remove
 
> Yes, it does. If secret is the same for all numbers then val is unique if
> seq_no is. Encryption is a one-one and onto map.

Very true. I started with the pain of managing millions of existing
outputs that I needed to sort and weed duplicates from and was thinking of a
more manageable solution. Then I completely missed how simple it was...

Cheers,

Paul