Re: JavaScrypt Encryption & Decryption - Codegroup bias?
- From: roger_for_nntp@xxxxxxxxxxx
- Date: 26 Apr 2006 19:52:12 -0700
frogstar wrote:
I've noticed that the alphabetical groups produced by JavaScrypt
do not contain the letters "Y" and "Z". Won't this weaken the
key as it introduces a bias into the system, and might be
exploited by an attacker?
Most modern computer based encryption methods have randomly distributed
binary outputs, which can be difficult to transmit or display.
Consequently it is a common practice to wrap the output in some
intermediate format which is easier to display. This format has nothing
to do with security, it is purely for manageability and convenience.
The "JavaScrypt" Javascript encryption libraries offer three such
formats, of which two (hexadecimal and Base-64) are widely
standardised. The third format, "code group" is something they made up
themselves and (as it says in its internal comments) the design is
based largely on aesthetics, i.e. they think it looks good. The absence
of Y and Z in the output is an artefact of this design. It does not
affect security because it is just a reversible representation format,
any attacker would reverse it to get the raw inary output before
starting an attack.
.
- Prev by Date: Re: Implementing byte stream cipher
- Next by Date: Re: Tiny table AES implementation
- Previous by thread: Implementing byte stream cipher
- Next by thread: is it sufficient to solve factoring problem
- Index(es):
Relevant Pages
|
|
