Y.K.: I think that your article is not a proof that evolution of individual technical systems is unpredictable but rather a possible explanation of why TRIZniks failed on predictions. You listed other possible explanations yourself: Altshuller laws are flawed, etc.
G.F.: You're right. I didn't mean it to be a proof, but to express my view. It proves a negative statement, though: laws of blah-blah evolution don't necessarily lead to a "guided" evolution.
Y.K.: You compare Laws of Technical Systems Evolution (or LOTSE, as you call them) to the gas laws. But why not to compare them to the laws of mechanics? Mechanics is concerned with laws that govern individual bodies. Thermodynamics is concerned with laws that govern the multitudes of bodies (molecules). Thus, one should ask a question, which category LOTSE belong to: to the category of laws that describe the behavior of individual bodies or to the category of laws that describe the behavior of large assemblies of bodies. It seems to me that from its name "Laws of technical systems evolution" Altshuller meant them to belong to the "individual bodies" group. Otherwise he would call them "the laws of technosphere (i.e. the set of all technical systems) evolution".
G.F.: Of course, one can compare LOTSE to Newton laws. One even can claim that no comparison can be made and these are unique laws, different from anything else. I mention a couple of reasons that make me think that LOTSE are of the "gas laws" kind. One, LOTSE are the result of noticing repetitive occurrences among large quantities of inventions. The other, evolution is viewed as a long-term, multi-step process. What and why to compare to, would be a separate subject, next step, which can be made only after the first, which is this article. Let readers ask questions and bring arguments, and then I'll go to the next step. Or, may be, I'll leave both (three, four...) views open. Can anybody refute my view?
I don't know what Altshuller meant. It seems to me, he meant different things in different contexts. Anyway, this has exclusively historical significance. The forms he put the laws into, refer to sets of inventions rather than to individual inventions. This is quite obvious regarding S-curves. But it holds for other laws, too. For example, "systems change from macro to micro", what does it mean? All systems always change from macro to micro? All systems sometimes change from macro to micro? Some systems sometimes change from macro to micro? Each system at least once changes from macro to micro? Etc. Since the law does not define a condition, all it says is, among a large set of inventions changes from macro to micro repeatedly occur.
Y.K.: To me the following is correct: systems may change from macro to micro. They may also change to something else. That is why I proposed a concept of the trees of evolution. But it is still individual evolution, not evolution of large assemblies as a whole.
G.F.: Yes, it is about an individual system - only as one of many. "Systems may change from macro to micro" means that in a set of systems some do. It's a probabilistic statement, which refers to a set. A tossed coin may land heads up. It also may land tails up. In a set of tossed coins some do the first and some, the second. These says nothing about what will happen next time I toss a coin.
Contrary to this, Newton laws refer to an individual body itself.
Y.K.: I am not sure that there are any probabilities in my concept, but there definitely are possibilities. (As is known from the theory of probabilities, not for any possibilities one can consistently assign probabilities, i.e. numbers from 0 to 1.) Still, the trees of evolution describe the possible changes of individual system, not large assembly of systems as a whole. On the contrary, you compared LOTSE to the gas laws, which say nothing about behavior of individual molecules. From the gas laws one cannot infer behavior of an individual molecule at all, neither deterministic nor probabilistic. May be your analogy was simply incorrect and you meant to say that LOTSE have to be probabilistic rather than deterministic. If it is the case, then I almost agree. Almost, because I believe that they even have to be rather "possibilistic" than probabilistic.
(The readers are welcome to participate in the discussion by asking their questions and presenting their arguments)