Serious about cryptanalysis? Learn a programming language!

Cryptanalytic program written in Perl

I’ve been a computer programmer since the early 70’s, about the time I became enamoured with cryptanalysis.  Writing computer programs to aid my cryptanalytic research has been invaluable to me throughout this entire period.  Sure, I have spent delectable hours solving ACA-type cryptograms by hand.  When I worked on more serious ciphers in university, however, computers have always been invaluable and time-saving.

My first cryptanalytic programs were written in PL/1 and ran on IBM mainframes.  The speed of these babies matched that of modern day PCs.  In the 80’s I “graduated” to PCs running compiled C code and the occasional AWK scripting program to help analyze and solve more sophisticated ciphers.

For many years my cryptanalytic scripting language of choice was AWK.  When all I wanted to do was parse the ciphertext, perform a frequency count, and perform some complex analysis, a scripting language is light enough for the classic edit-run-debug loop.

A few years later I moved on to Perl.  Here was a scripting language that let me do anything I pleased in a minimal amount of coding time.  My research output increased significantly and I was able to produce important results I could never have done by hand.

An example of results I could never have done manually is the initial work I did on John F. Byrne’s Chaocipher.  Sifting through the 13,300+ characters of Exhibit 1, looking for pt/ct identities, plaintext and ciphertext double letters, isormorphs, trawling a pattern through the plaintext or ciphertext looking for a match, performing statistical calculations, simulating alternative cipher models and enciphering the 13,300+ plaintext characters — these are only a few of the tasks that scream out for automation using computers.

Recently I spent time trying to solve a single ADFGVX ciphertext challenge message.  The tremendous amount of tedious and error-prone computations and calculations needed to match columns would take days to perform, so I wrote some help program (in Perl) that let me concentrate on the analysis without getting bogged down in the “drudgery”.

Up-and-coming cryptanalytic techniques we’re seeing more and more are Hill-Climbing and Simulated Annealing (SA).  You can only take advantage of these powerful methods if you can write a program to do it for you.

You don’t need to study Advanced Computer Science to be able to write computer programs.  Pick a language (e.g., Perl, C, C++, Ruby, Python, Visual Basic, Java) and start writing simple programs: reading text from a file, taking a frequency count, writing to standard output, manipulating characters and strings, loops, and performing mathematical calculations.  In a short time you will have mastered everything you need to write cryptanalytic aids.

Don’t worry if you don’t know the difference between a double-linked list and a B-tree.  Just program the method you would use if you were solving it by hand.  This alone will cut your solving time by 1-2 magnitudes.

In my opinion, every cryptanalyst should know how to write computer programs to ease the work.  And I believe anyone can acquire that skill.

Don’t lose out on advancing in the world of Classical Cryptanalysis because the problem key space is too large for manual solution and you can’t program.  Take that step, come on into the modern world, and let your cryptanalytic achievements soar.