[Hardware] "success"

[Dan Oetting]

On Oct 18, 2006, at 5:26 PM, Jim C. Nasby wrote:

> Well, one of those rules that would have to change is stats, since
> that's a prime motivator for people to cheat. I don't see that  
> working,
> though.

This is trivial to solve. All you have to do is change what you count  
for a completed block of work. Currently you count the claim that a  
client completed a particular work unit of 2^28 keys. If you simply  
change the stats counter to finding a key that correctly generates  
the first 28 bits of the cypher text, the work can be easily verified  
so there will be no chance to cheat by short circuiting the work  
(changing to 32 bits and counting each as 4 blocks would simplify the  
client). You also get a free verification that the client is finding  
and reporting cypher text matches which is something that is missing  
in the current clients.

-- Dan Oetting