I see three problems with the Internet socionics community right now. Two of these problems have to do with type descriptions, and one has to do with testing.
Problems with type descriptions:
1. There are inconsistencies in the descriptions. I can see this when I look at the ILE description. www.socionics.com basically describes the ILE as someone who likes generating ideas and is somewhat absent-minded. The description at www.socionics.us is very similar. However, the descriptions on this site and on www.wikisocion.org describe the ILE as more enterprising and being oriented towards what's interesting and what isn't. I understand that there are going to be different interpretations of the various types. However, these descriptions seem different enough that it could cause confusion in someone trying to identify their type.
2. The logical and intuitive types are described more from a skill-based standpoint, while the other types are described more based on emotions or strictly personality-based characteristics. I suppose that this is okay, but when www.socionics.com describes the ISFp as someone who is "never the best or the worst", I can't help but wonder if this is turning people off. Whether you like the book or not, I think that "Gifts Differing" does a good job of describing the skills of the various MBTI types, so that one isn't left with the feeling that one is, for example, only emotional and there isn't much else to them. The skills and talents of every type should be described. In Gifts Differing, types are described at being good at everything from science to working with machinery to woodworking. If socionics wants to appeal to people, it should take a similar approach. Simply describing someone as emotional, friendly and that's it doesn't cut it.
Problems with testing:
There are no good socionics tests out there. I think a good approach would be a test like this: http://www.cognitiveprocesses.com/as...nt/survey.html. The questions would be based on the socionics theory of functions, and the highest scoring function coupled with the next highest scoring compatible function could help someone discover their type.