Since the three original topics got the life of their own, I'd like to resume what I have learned there, and to refresh the still open questions.
It seems to me that the consensus regarding evolution according to TRIZ is as follows.
"Evolution from A to B" means that
D1) A and B were used for the same purpose P, and
D2) B replaced A.
Q1. Are the conditions D1 and D2 sufficient, i.e. does ANY pair of systems A and B, which satisfies D1 and D2, constitute "evolution from A to B"?
Q2. Are the conditions D1 and D2 necessary, i.e. does a pair A and B exist, which does not satisfy D1 or D2, but still constitutes "evolution from A to B"?
Q3. What other types of evolution, which cannot be described as "evolution from A to B", exist in TRIZ? What do they mean?
Q4. Regarding the TRIZ laws of systems evolution: What do these laws mean? Do they say how systems evolve, or rather how systems should evolve? In other words, are they scientific laws describing how things naturally happen, or rather engineering laws describing how things should be done?