ACNews #12 07 октября 2002

Программирование

Docs - encryption and decryption of messages.

■ ■ ■ ■ ■ ■ ■ ■ ■ ■

  ■ ■ ■ ■ ■ ■ ■ ■ ■ ■ ■

I have long been tempted to build at least one of the plugins 
though a bit of intelligence, well, oh you do not want to fix 
yourself a letter after the conversion of another letter from 
another inetovskoy encoding! How many of them out there, these 
encodings? God knows. some written in transliteration, but 
whether it is good, especially when you consider that Laws 
change letters every individual and their original? (I 
checked:)) think it's easy to teach the program to decipher 
this debris from the letters, covering the owner's mistake?;) I 
make an intelligent converter with a cunning system of rules, 
but not today I will write it and did not publish today, these 
rules you:) Today I will write about the problems of encoding 
"BUG", had taken out nowhere and spoil half the letters in my 
inbox ... 

BUG: I can not draw you a table of characters. Even part-time! I
it just was not:) plugin written by an educated bet,
until the text was transformed into something meaningful, I'm 
in my childhood in such games played, for example ... 

-------------------------- ----------------------- Encryption 
---- Well, who in childhood is not encrypted messages? when the 
message is high, and each letter corresponds uniquely a symbol, 
it is enough frequency analysis to unravel the meaning of the 
text. and if the encryption of the information is destroyed? 
you foreign phones with the letters seen? for each digit - of a 
few letters, but they are not optimally placed, all groups 3, 
except for two separate groups ... message printed the phone 
numbers that uniquely difficult to decipher ... I came up with 
placing a child of Russian letters on the numbers, though also 
maintaining an alphabetical order, but with a more reliable 
decoding - almost any word ("not even the phrase!) can be 
solved uniquely and almost algorithmically:)



                  1 2 3 4 5 6 7 8 9 0

                  and it is given on p y x y

                  b r g l s n a r e tsch

                     q mn tonnes shsch th

                                             I
that is interesting, and itself is placing letters regularly, 
just try to find this pattern:) 
-------------------------------------------------- 
------------- Here and in our encoding all was not well with 
the unique data recovery. lost just one letter, but what ... 
rule: in the FIDO letter code # 8d can not exist, and like it 
or not, and turn in either Latin H, or a transfer Line ...

clarification: the CP866 # 8d - Russian H, Win1251 - not used
in KOI-8R - not used in ISO - pseudo-character,
BUG - on Russian.
corollary: the encoding BUG turns of Russian in H, and discern 
where it was H, and where on - definitely not. up a table where 
the vertical postpone all variants of the symbol To the left of 
incomprehensible letters H, and horizontally - all variants of 
the symbol on her right, the cells inscribe the most logical in 
this If the value of the letter H.

(NaL - not a letter)


              after az AZ a-z A-Z H NaL

              to


               AZ of?? ?? ?? oh oh

               AZ of Nrus?? ?? Nrus? about?

               a-z??? ?? H??? ?? ??

               A-Z??? ?? H H H H

                H H H on Nrus on about?

               NaL at? Nrus H H Nrus? about?

see how vague and ambiguous options? therefore, envelopes, the 
envelope is bad, that is: ( may be worthwhile to write a 
plug-seeker and replacing Latin H and p in Russian, where do 
you want? By the way, expressed gratitude to Alex Nikiforov, 
who sent I encoded BUG extensive text Necromancers, which 
allowed more fine tuning and testing of plug-in: