Posts by Tommy

1) Message boards : Number crunching : method question! (Message 380)
Posted 23 Dec 2007 by Profile Tommy
Post:
It's definitely a tough one - previous two were broken in just few weeks, this one took already lots of CPU time and it's still unbroken. Maybe its an 'offizier' (double encrypted) message, which probably couldn't be broken with current software, or it's too garbled or has too many missing letters to decrypt. We won't know until it's broken.


In some cases, the Germans using double-encrypted messages (offizier). I read that transformed the message with a specific codebook and then crypted with enigma. So with the current client is impossible to decrypt a message double-encrypted :(. This case will have a particularly important content :).
2) Message boards : Number crunching : method question! (Message 366)
Posted 16 Dec 2007 by Profile Tommy
Post:
What is the "criticality" of this message that we are trying to decrypt? It seems to me that this is harder to solve. Thanks
3) Message boards : Number crunching : Realtime graphs (Message 359)
Posted 14 Dec 2007 by Profile Tommy
Post:
We should have graphics with longer periods because otherwise are almost constant.
4) Message boards : Number crunching : logo (Message 356)
Posted 14 Dec 2007 by Profile Tommy
Post:


Small, plain and by far not as colorful as the other attempts, but easy on the eye :)
Fits better into a template with dark background and light font colors though.

Edit:

Slightly enlarged version (120%)



Very nice ... however, we should also change the letters are repeated (for example the "m").
5) Message boards : Number crunching : method question! (Message 355)
Posted 14 Dec 2007 by Profile Tommy
Post:
I read that the maximum score achieved is 1883763. It is possible to know "a priori" (an estimate) what will be the score of a decrypted message of a certain length? You can make a statistical average of messages already decripted so that we can predict the score for messages not yet clear?
6) Message boards : Number crunching : method question! (Message 269)
Posted 22 Nov 2007 by Profile Tommy
Post:
A score tells how well the result resembles plaintext. The final score reported to the server is the best result found by your client within work unit range.
It's based on trigram count - if you take a look at 00trigr.naval, you'll find groups of three letters, each group has it score which depends on how often it appears in decrypted enigma messages.

If you want to know more about hillclimbing enigma messages, you can find interesting article here : http://frode.home.cern.ch/frode/crypto/bgac/index.html


Hi TJM, there is a mistake: this post was opened by me and not by Alfredo15. The first post was written by me (Tommy: p)!
Then i wanted to get a question ... the server status page... Hceyz72 work units ready: 5353552 ... the total number of wus? After the five million wus will certainly decrypted the message? For now we are about 350000 wus calculated, this means that we have analyzed the 15% of the total, right? Thanks!
7) Message boards : Number crunching : method question! (Message 142)
Posted 14 Nov 2007 by Profile Tommy
Post:
I wanted to know as the program tries the solution of the enigma message. The project uses a brute force method, trying all the possible combinations? As an example, what indicates the score in the solutions? Thanks.






Copyright © 2024 TJM