Separate words into vectors of tuples

[modass]

For the algorithm, we focus either on the sets of TA-tuples (reg_A to define C-Classes) or on ATA-tuples (to establish the domination order, since for the TA-components this is equivalence). Thus, we can store those two vectors individually to increase efficiency.

Details:
The monotone domination order induced by the subset-relation for the TA-part of a word is equivalence (with respect to the TA-components). The number of TA-tuples in a word is fixed and finite, only the number of ATA-tuples may vary. Thus, with respect to the TA-tuples, two words w1, w2 can only be related via the subset relation w1 subset w2, if for each set in the word w1 there is the equivalent tuple in w2 - since there are only finitely many, this yields equivalence.