1. Socionics started and continues as a MODEL approach which models information processing of the psyche. The generally accepted model now is the Model A. If we forget this, then we completely miss the point. If we think that socionics studies real people in real life(instead of models of TIMs), or types as described in some descriptions, or relationships between real people (instead of relations between TIMs) then we are missing the point.
2. Typing in socionics is nothing but TIM IDENTIFICATION based on the model A. If you have demonstrated correlation between person’s behavior or thinking or acting and the respective Model A, then and only then you have identified this person’s type. If it seems to you that the person looks similar to someone who you consider to be such-and-such type or if he seems to conform to some description or if he seems to be in a certain relationship to a “known” type or if he has passed some test or else, then, please have no illusion - you still didNOT identify his type.
3. RELATIONS in socionics are NOT descriptions of some real life observations but analysis of interactions between different TIM modelswhich takes in consideration interactions between the models' functions that process the same information elements.

Basically that’s all. But it has very serious consequences:

Even if you think that talking about types is enough to be doing socionics, you might be wrong. A “type” does not necessarily mean a “model” – there are plenty of types without any model behind them at all. Take for example the temperaments or the Myers-Briggs types. They are just some type names with some description behind them, whereas a true model would be a set of elements with defined relations between them.

If you are typing by using some tools like the Reinin’s dichotomies or Myers-Briggs indicator or VI traits or DNA analysis or brain MRI or fingerprints or handwriting or horoscopes or numerology or any fancy stuff like that, please, take care to demonstrate (either theoretically or practically) that there is a definite correlation between those tools and the Model A. Otherwise you might be well typing some “other” types.

Use of socionics in order to describe some real life situations without referring to specific models or mechanisms of model interactions could be very deceiving. If you did not identify (or supposed) the TIMs of the participants of a real life situation and if you do not consider interaction between their functions how can you be sure that you are making a socionics analysis at all?

Providing oneself as an example of such-and-such type’s behavior could be very incorrect mainly for two reasons: (I) one cannot judge himselfadequately when it comes to his low-dimensional (you can also call them weak/vulnerable) functions; (II) one’s real life function’s acquired experiencecan be different. For the same reason could be erroneous an argument of the type: “I would say/do the same thing, hence he/she is my identical type”.

As far as one realizes all the said above, I personally don’t mind if one is using Model A or not. If one is using numerology or astrology, or any other typology, please, keep in mind that this is NOT socionics.