Entropia Project
god doesn't play dice with the universe
A german paper about the forseen methods for the analysis is in progress. This is a small summary of the used techniques. Some of them were already mentioned in de.sci.mathematik as 6891c$464beff0$54480012$13001@news.hispeed.ch (05/17/2007).

Length Analysis
First of all a length analysis of the values is realized. This means the minimal, maximal and average length of the values in bytes is determined. Usually the random values have the same length (e.g. TAN have 4 chars, MD5 has 32 bytes in hexadecimal, etc.).

Charset Analysis
All symbols used within all the random values are written into an array, This one defines the charset which is used within the value listing (e.g. mostly hexadecimal for common hash algorithms as like MD5 or SHA1). This makes it possible to create a fingerprint for the values, which will ease further bruteforce attacks.

Entropy Analysis
A statistical analysis is going to identify the entropy of the values. This is one of the most greedy tests because the entropy of single values, blocks and the whole listing is analyzed.

Redundancy Check
Regarding the charset determined a redundancy check is initiated. This one is going to identify, if the same char could be used within a value or a list of values. If no redundancy could be detected, a permutation is expected. Especially self-written algorithms come up with such limitations.

Object Distancy Analysis
Every symbol within a code is replaced by a number. This makes it possible to calculate the arithmetic distance between values and blocks of values. Incrementation and an even distribution of characters can be determined this way.

Inherit Path Analysis
Every symbol with the charset is highlighted within the value listing. This makes it possible to create a path diagram which is able to detect the common positions for symbols and sequences. Weak algorithms, usually using array and shifting techniques, could be determined with this test.

Bibliography
Simard, R., L'Ecuyer, P. (2007), TestU01 - Empirical Testing of Random Number Generators, http://www.iro.umontreal.ca/~lecuyer/myftp/papers/testu01.pdf

Undocumented Results

md4Symbol FrequencySymbol AverageSymbol CumulationEntropy StructurePath Analysis for a
md5Symbol FrequencySymbol AverageSymbol CumulationEntropy StructurePath Analysis for a
sha1Symbol FrequencySymbol AverageSymbol CumulationEntropy StructurePath Analysis for a
chitSymbol FrequencySymbol AverageSymbol CumulationEntropy StructurePath Analysis for a
crc32Symbol FrequencySymbol AverageSymbol CumulationEntropy StructurePath Analysis for a