Here are some things that we need to do away with if we are to accurately type people:
3. Liberal allocation of the shadow functions.
If a claimed INFJ seems more like an ISTP, people will just cite tertiary Ti and inferior Se as the culprits. If an claimed INTP seems more like an ISFJ, people will just cite tertiary Si and inferior Fe as the culprits. Sigh... the shadow does NOT work that way. A person's basic behavior while not under stress, would not be more like that of their shadow, at least not until they're older and they've consciously tried hard to integrate it. I can tolerate a provision for allowing the shadow traits to come up once in a while when it's not part of the typical pattern, but you can't use that to dismiss a huge, major pattern in their personality!
4. The T/F gender socialization card.
I can buy it only in very limited circumstances. Sure, this might explain away a few superficial traits, or maybe even major ones if the person had a VERY harsh upbringing that emphasized gender roles and was forced against their preference repeatedly. I don't think this is as common as people are imagining, though. The reality is that this card is effectively used to make it such that a woman can show far more F preferences and still be called T, while a man can show far more T preferences and still be called F. They're basically lowering the standards in an attempt to make the distribution of T and F more equal. But in reality, it doesn't work that way. The standards should not be lowered, because there are T women and F men who have CLEAR preferences, and often without even going too far outside of gender roles. Accept that this function sometimes shows a gender bias and move on... don't try to artificially "repair" that bias with this card, please? You're only making it harder for women that are only slightly expressed Fs, or men that are only slightly expressed Ts, to determine their type.
6. "Weak" letters that don't fit the type pattern.
One's dominant function should never be "weak." One can have a weak auxiliary, but the dominant function should always be fairly strong, or that means you're a different type. I would think this would be obvious, considering that the dominant function is supposed to represent the consciousness and the self-image.I can accept a weak I/E, because that just means you have a strong auxiliary. I can also accept a weak auxiliary indicated by a weak letter associated with it. But I do not buy the whole concept of a weak dominant, or a weak J/P. J/P has less to do with behavior in terms of organization/punctuality/neatness, and more to do with functions. The difference between J/P types is a completely different functional order, and you should be able to tell which one you have, unless you're completely mistyped.
8. Assumption that functions can be individually developed.
This doesn't work. When one has Fe, they have Ti. They can't go out and develop Te. Sure, they can emulate Te behaviors and learn to communicate effectively with Te users, but that's not developing Te. They may be learning skills traditionally associated with Te, but their motivations for learning them will likely be related their dominant or auxiliary. Fe would want to learn Te skills in order to get along with Te users, for instance. Ni might want to learn Te skills in order to more effectively shape their environment into what they would like to see. Give me just about any skill or concern, and I can give you each function's possible motivation for having it.