Discussion of the article "The Mistaken Macro-Micro Evolution"

Y. Karasik: Transition from a Macro- to a Micro-level is mentioned in TRIZ in two contexts: in the context of the laws of evolution, and in the context of the Standards. In the context of the laws it looks as follows: "Evolution (or development) of the working bodies of a system initially unfolds at a macro-level and then at a micro-level." In the context of Standard #3.2.1 it looks as follows: "Efficiency of a system (at any stage of its evolution) can be increased with the help of transition from a macro- to a micro-level. In the course of such transition the system (or its portion) gets substituted by a piece of material capable of performing the same action as the whole system (or its portion) when interacts with a field."

Your article deals with the law. As for the Standard, I have to say that its purpose was not to claim that development of the working bodies first unfolds at a macro-level and then at a micro-level, but to say that implementation of a function at a micro-level is ALWAYS (or OFTEN) better than its implementation at a macro-level. I know that because I was a co-author of the Standard and the author of its first wording, and my idea was like this. By the way, the assumption that micro implementation (when progress of science makes it available, as you noticed) is always (or often) better than macro implementation is not so obvious.

G. Filkovsky: I disagree that implementation of a function at a micro-level is ALWAYS (or OFTEN) better than its implementation at a macro-level. Let's assume that there really are many more good inventions implementing the macro-micro transition, than inventions implementing a micro-macro transition, or the last even does not exist. It occurred to me, that this fact would not necessarily point to an inherent advantage of micro-solutions compared to macro-solutions. It could be simply a statistical effect. Let me explain using simplified assumptions.

Let's assume that on average macro-solutions are as good as micro-solutions. Let's say we have random 100 problems, each of which has a possible macro-solution and a possible micro-solution. Because of the assumption of "equal average" above, 50 of them have a good macro-solution and a bad micro-solution, and the other 50 have a good micro-solution and a bad macro-solution.

Now, let's start in the moment of time that no micro phenomena are known yet. Inventors use only macro-solutions for all 100 problems. 50 problems have good solutions then, and other 50 have bad ones. Some time later, half of micro phenomena are discovered. The solutions of the first 50 problems do not change, since they are better than their micro alternatives. Half of the other 50 solutions switch to the newly discovered micro-solutions. Now we have 75 good solutions and 25 bad ones, thanks to the macro-micro transition. Some time later, the other half of micro phenomena are discovered. The rest 25 solutions switch to micro. The macro-micro transition brings up 100% of good solutions.

And this happened not because micro-solutions are always, or even often, better than macro-solutions, but only because the domain of available solutions widened. This exercise shows, in my view, that the "macro-micro transition" is a good heuristic for engineer to remember, but it has nothing to do with a nature of macro vs. micro implementation. There is a long list of systems, which stay on macro-level. It is not possible to determine, if they stay macro because an appropriate micro phenomenon is unknown yet, or rather because they are the systems for which the macro implementation is better than a micro one.

Y. Karasik: How does this refute the hypothesis that "Micro implementation of a function is always better than its macro implementation" ?

G. Filkovsky: This does not refute that hypothesis, but rather shows that the same facts can be explained without it. Any idea for a critical test?

Y. Karasik: Which facts ? The facts that some old macro implementations get substituted by new micro implementations ? The hypothesis was neither based on such facts nor purported to explain them.

The hypothesis was proposed in order to give inventors a heuristic: "if you have an option of implementing something at both macro and micro levels, choose the latter. It will likely be better." That's all.

G. Filkovsky: This is a very good clarification. It emphasizes a difference between macro-micro heuristic vs. an (inherent) pattern/law of systems evolution.