As time has gone by, my understanding of socionics has become both deeper and more tangible; I have developed my own typing system, and various examples of the the information elements have become clear to me. However, very recently, I have realized that some people have taken socionics too far; they develop their own typing system, and believe that everyone neatly falls into one of 16 types - with all relations in place (at one time, I had, in fact, been one of these people). I have started to deviate from this. I believe that there are indeed variations within the individual types that even show themselves in intertype relations; not all LIIs identify with each other and relate well. Because of this, I have been looking into extensions of socionics. I have looked at some of the major socionist's ideas. The main problem I have is that many of these ideas tend to reuse notions of information metabolism (e.g., the DCNH model). While these systems are not completely contrived, I would think that, based on sheer notions of probability, the best extensions would come from outside of socionics. I think that the tendency for extension through elements of information metabolism is sort of holding the field back. Perhaps this is the wrong audience, but what I am trying to say is that if any of you are seriously thinking of extending socionics, I would suggest that you try and consider outside ideas.