When I was studying in AzOIIT from 1973 till 1975, Altshuller taught us ARIZ-71. (It was actually its latest version called ARIZ-71B.) I believe it was the best version of ARIZ ever.
The students were instructed to start solving problems by ARIZ not with step 1 but with step 2.3 and continue until step 3.8 only ! Probably because only these steps had a clear logic:
Never ever solution proceeded further than step 3.8 or started before step 2.3. But out of curiosity I familiarized myself with other steps. Their logic was already much less clear. For example, resolution of physical contradiction was followed by formulating a technical contradiction between the solution and the prototype ! What for ? In order to further improve the solution by resolving the technical contradiction with the help of the contradiction matrix ? A dubious objective. That is probably why Altshuller never required students to follow through these steps.
How, nevertheless, this logic came about ? The answer to this question can be found through the analysis of evolution of ARIZ.
In the first ARIZ, namely in ARIZ-56, the search for solution used to begin with formulating a technical contradiction. The notion of IFR and physical contradictions did not exist yet. Basically ARIZ-56 boiled down to formulating a technical contradiction and its resolution. The technical contradiction was actually the contradiction between various prototypes: one prototype is better than another in A but worse in B.
ARIZ-59 already features the notion of IFR. The technical contradiction becomes not a contradiction between various prototypes but a contradiction between a selected prototype and IFR !
ARIZ-61 is a further improvement of ARIZ-59 in the number of ways technical contradictions could be resolved. It has to be said that contradiction matrix did not exist yet. Instead the bulk of steps consisted of applying the prototypes of the 40 principles one after another. The contradiction matrix was first introduced in ARIZ-65
ARIZ-69 already features the contemporary contradiction matrix. But in the rest the logic remains the same: formulate IFR first and then technical contradiction between IFR and the most promising prototype.
Formulating a technical contradiction between IFR and a prototype is not as easy as formulating a technical contradiction between two prototypes. Actually in most cases it is impossible to formulate a technical contradiction between IFR and the prototype at all. That is why ARIZ-71 replaced formulating such a technical contradiction by formulating a physical contradiction between IFR and what does not allow one to achieve it. But formulating a technical contradiction did not disappear from the algorithm. It was simply shifted to after physical contradiction is resolved. A dubious motivation was provided for keeping this step in the algorithm.
ARIZ-77 combined ARIZ-71, SuField Analysis, The Standards, The Lines of Evolutions, etc. into one algorithm. It was hard to find a place for technical contradictions in it. And Altshuller again moved them to the beginning of the algorithm: technical contradiction first, then IFR, then physical contradiction, then SuField, etc. ARIZ-77 started to simply resemble the history of TRIZ. The sequence in which various elements of TRIZ are encountered in ARIZ-77 reflects the chronology of their invention: technical contradictions came into existence first, followed by IFR, followed by physical contradictions, etc.