Re: AES as hash function and PRNG

Pécuchet wrote:
> Volker Hetzer <volker.hetzer@ieee.org> wrote in
> news:d8msq5$6ge$2@nntp.fujitsu-siemens.com:
>
>
>>>Any problem with changing that to:
>>>
>>> ...
>>> 5 = 1
>>> 6 = 0
>>>
>>>so that you always get at least one bit per poll?
>>
>>Yes, then 01 and 00 happen twice as often as the other values.
>>
>>Lots of Greetings!
>>Volker
>
>
> I missed the original post but, from what I understand, 1 to 4 rolls
> produce 2 bits of random data whereas 5 and 6 rolls produce only one bit of
> random data. I don't see any problem with that. There is no bias.
Ah, ok. I thought it was a spelling mistake.

Thanks for the correction!
Volker
Received on Thu Sep 29 21:44:19 2005