This is what is expected in a continuum. If in fact AD, HFA, AD, etc are all part of the same spectrum, then any criteria used to divide them are going to be rather arbitrary. Sure, one can distinguish a classic Kanner's person from an Aspergian, for the most part. When you get to the 'higher functioning' people diagnosed with AD (or the 'lower functioning' people diagnosed with AS), the distinction becomes very hard to make.
It's rather like height. We can all agree that if you are seven feet, you are tall. We can all agree that if you are four feet, you are short. Between those extremes, it's all rather fuzzy.
[edit] I just noticed that Geek made my point far more eloquently than I did. And with citations. Good on you.