It scans a file listing a succession of states for a given set of variables. It tries to find what causes a given variable to switch to a given state.
This variable and this state are defined by the user (i.e. they are inputs to the program) and are called the desired variable and desired state.
 

Two pieces of information are sought by toukon:
 

• a trigger variable that switches simultaneously with the desired variable.

• the initial conditions that allowed the trigger to switch the desired variable to the desired state.

For example, consider that the input file contains the following:
 
 

If you ask toukon to find the reason why the "desired variable" switches to green, it will answer that it's because "variable 1" switches to green when "variable 2" and "variable 3" are initially green.

Its output would be :

Trigger:
Index = 0
State = green

Initial conditions:
State #0 is NULL
State #1 is green
State #2 is green
State #3 is NULL

The overall project is called toukon. But it is actually made of two sub-modules:
 

•  anumaana (means inference in sanscrit): looks for the variable that switches simultaneously with the desired variable the most often. This variable is considered to be the trigger variable.

•  upamaana (means comparaison in sanscrit) : separates the occurences when the trigger and desired variables switched together from the occurences when the trigger switched without the desired variable. Derives the combination of states that are the most common to the first set and never appear in the second set. This combination is considered to be the initial conditions.