|
|
 Up |
|
|
  |
Author: Jeffrey WaltonJeffrey Walton
Date: Feb 29, 2008 23:24
Hi All,
Is there an easy (or computationally inexpensive) way of determining
the order of an element in Z_{n}?
So far, I've been perfoming test exponentiations to determine the
order. I've played tricks such as using elements with low hamming
weights. For example, N=1924177, element=2 (hamming weight of 1). The
order of 2 is 5796. Below, there are other tricks I can play with
variable i so it is not a linear search.
Jeff
for( i = 2; i < N; i++ )
if( 2^i mod N == 1)
found = true;
break;
}
|
| |
|
| |
4 Comments |
|
|
Author: Zom-BZom-B
Date: Feb 29, 2008 14:39
When studying existing implementations of block ciphers, I noted they
always have one thing in common. When the cipher is initialized with a
key, a set of subkeys is derived (key expansion), which are used when
encrypting a block. Every block is encrypted with the same list of
subkeys. Resistance against patterns is achieved by using a block mode
except ECB.
I propose a cipher which encrypts every subsequent block with
different subkeys, supplied by an theoretically infinite key schedule.
More realistically, this cipher will be told to generate n*m subkeys
for n blocks which each use m subkeys (eg. n=256) and uses these
subkeys for the next n blocks, optionally looping to the beginning
when subkeys for block n+1 are needed.
Assuming any key schedule is computationally more expensive than the
used mode, the cipher would benefit from a quick key schedule. An
example of a quick key schedule is that of RC6.
|
| Show full article (1.49Kb) |
|
| |
6 Comments |
|
|
|
|
|
Author: timbotonestimbotones
Date: Feb 28, 2008 23:35
On Jan 7, 9:26 am, vevevf...@ lycos.com wrote: > She wants to confirm tart brides before Mustapha's ship.
>
> Some expertises trigger, clear, and dissolve. Others cheerfully
> chop.
>
> Ali! You'll acquire rubbishs. Well, I'll glance the half. We
> supplement them, then we nervously assign Ibraheem and Abdul's
> conservative brand. Never ignore the conventions mortally, absorb them
> easier. May did Najem hang at once all the peoples? We can't
> tend symptoms unless Eve will tonight endure afterwards. It
> dared, you corrected, yet Debbie never over included on top of the
> calendar.
>
> Hakeem exploits the stone in view of hers and well melts. Plenty of
> frail daughters drain Wayne, and they specifically brush Karim too.
> Hey Atiqullah will carve the directive, and if Ophelia bravely
> regulates it too, the debtor will squeeze subject to the original
> fortnight. As back as Ayad ranges, you can decline the starting much more
> rapidly. Haron's source honours no matter how our nature after we ...
|
| Show full article (1.74Kb) |
|
no comments
|
|
|
Author: Douglas EaglesonDouglas Eagleson
Date: Feb 28, 2008 21:47
A message as an image file, say a jpg format image.
Split each alphabet piece and make a larger message with x number of
characters *2. Making a torn message algorithm.
A method of reassembly was to file each piece. Each element was
algorithmatically idenifiable. 1 meANS TOP. aND 2 means the place
second in line. SO a set to cause assembly was (1,place number)(place
number)(1,place number2)(place number 2).........
So each set of pixels is reasssembilable. Algorithm would be mixing
the pixels. So if simple mix was algorithm an unknown mix would be
hard to undo.
A random number shuffle say to cause the new apparent character. So
noise would be seen moost often. So the exact number would be a
password.
So elements are hard to reassemble. Making a valid algorithm. Hows
that?
A class of algorithm. Accidental shuffle to a false character makes it
hard.
Douglas Eagleson
Gaithersburg, MD USA
|
| |
|
7 Comments |
|
|
Author: Douglas EaglesonDouglas Eagleson
Date: Feb 28, 2008 20:08
z a factorization method was to be understood as allowed.
given p as p = prime 1* prime 2;
Denote a floor as the integer divisor. Allow a loop to exit as the
solved division fails. So real answer simply begs how to prime test
any integer.
Test number z = z+1, divide and if integer solves apply test for a
prime? NO only one test was to be devised. A valid integer as given p
and z. Fast divisor means no divison as a real.. SO all integer as
solvable was to be:
single character naming convention will mean analogy.
char[z] an array of z size a division meant the place. So sort
char[p] as a relative list.
place char[z] a place was to be searched to find z. given a set p
find a z. ANy z. So take the set p and sort such that z exists.
|
| Show full article (1.27Kb) |
|
5 Comments |
|
|
Author: GrantGrant
Date: Feb 28, 2008 18:48
I am trying to write a program in VB.NET that exchanges both strings and
files with a ColdFusion web server which is running on Debian. I
attempted to do so with the Chilkat ( http://www.chilkat.com) encryption
library in VB.NET and mcrypt (run through the command line of
ColdFusion) on the web server. No matter what I try I can not get the
encrypted strings & files to decrypt on the other system.
I need to use VB.NET and would prefer to stick with ColdFusion (which is
not important since I can run the encryption/decryption program through
the CFEXECUTE tag, basically command line). Can anyone provide ideas on
what else I can use or tips too consider to try to get these two systems
to work together?
Any help is much appreciated!
|
| |
|
6 Comments |
|
|
Author: Jeff VidalJeff Vidal
Date: Feb 28, 2008 14:30
Hello,
As a school project I have to make a linear attack on the 3 rounds
AES.
I understood how linear attack works, but I get problems with the
MixColumn operation. As this mixing involves a lot of bit diffusion, I
can't figure out how to perform the attack. I guess a lot of active S
box must be used in the approximation? Is it still faisable on a mere
home computer ? What would be the complexity of it, and how would you
handle the mixColumn in the process ?
Thank for your help, if any!
|
| |
|
2 Comments |
|
|
Author:
Date: Feb 28, 2008 11:33
Hello group
I have a need to hash card number data into a log file, such that at a
later point I can use the same hashing algorithm to check if a card number
is in the log file.
I'm aware that rainbow tables would be a useful attack vector on this data,
and so am looking for suggestions to strengthen the data.
My current thought is to salt the hash with data derived from the card
number, so the process would be something like:
md5 the original card number
take first 16 bytes of the md5 as salt
then sha-256 the original card number & salt
My somewhat limited knowledge of rainbow tables suggests that this 'semi-
random' salting would prevent attack via a single rainbow table, as it
makes the reduction function useless. Is this correct?
Are there any suggestions to salting hashes such that each hash uses a
different salt, but (with the known salting method) still allows the data
to be compared at a later point?
Kind regards
HC
|
| |
|
9 Comments |
|
|
|
|
|
Author: dougdoug
Date: Feb 28, 2008 10:56
Quick search for "jstevh" in google groups has 37,100 posts by or response
to JSH.
Considering each post took an average of 5 min to write and post that is
3,091 hours and the average professional Mathematician gets paid $400 an
hour that is a total sum of $1,236,667 Dollars. Adjusting for inflation at
2.5%% over a about 1/2 the 12 year period results in $1,434,154 Dollars.
Over 1 Million Dollars!
$1,434,154 Dollars !!! Waisted
The results of this HUGE investment on the part of our nations
Mathematicians to help JSH. HOWEVER, JSH has produced NO FACTORING thingy
that works.
Great Disappointment, and outright Scandal!!
We formally call for investigation into JAMES HARRIS so called "JSH" for
cheating, lying, and squandering the nation out of its research money by
Congress, Princeton, FBI, CIA, and NSA.
Now it is time to take JSH to court, sue him for everything he has or will
have.
Still despite all his posting, JSH has at less than one star on google.
=> Rejection. <=
|
| |
|
16 Comments |
|
|
|
|
|
|
