17 Dec
2003
17 Dec
'03
11:17 p.m.
A quick follow-up: I suppose a cut-and-choose protocol in the withdrawal would prevent this attack. Instead of sending in one blinded coin to be signed you'd send in 100 blinded candidates, then the bank would pick 99 and you'd reveal the r's for the others (remember, they are blinded with r^e) so the bank can verify they are of the proper form. The bank would then sign the one remaining one and return it to you. What a pain! I hope someone can come up with something better, or show that the attack doesn't work. Hal