[Cryptography] Real-world crypto/PRNG problem: Bridge

[Tom Mitchell]

On Mon, Aug 22, 2016 at 11:24 PM, Viktor Dukhovni <cryptography at dukhovni.org
> wrote:

> On Mon, Aug 22, 2016 at 06:17:31PM -0700, Jon Callas wrote:
>
> > Christian Huitema pointed out to me that 0x5DEECE66D is actually 36 bits
> > not 32. I used <http://www.binaryhexconverter
> .com/hex-to-decimal-converter>
> > to convert it to 25214903917,
>

The number of bits vs. the storage ?  All bits even 0's should or could
count.
ibase=16
obase=2
5DEECE66D
10111011110111011001110011001101101
$ echo 10111011110111011001110011001101101 | wc -c
36


So yes 36 bits.
But the first bit could be 0 then these tools would prune leading 0's
and we might say 34 bits.
Most commonly we would store it in bytes vs nibbles as 0x05DEECE66D for this
000010111011110111011001110011001101101  or with more leading zeros
depending on storage  (5 bytes is uncommon).  It is just 'luck' that the
left most bit
is 1 and gives us 36

ibase=16
obase=2
5DEECE66D
10111011110111011001110011001101101
00000005DEECE66D
10111011110111011001110011001101101
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.metzdowd.com/pipermail/cryptography/attachments/20160823/3f5f87d1/attachment.html>