Is the dot structure method failproof? in other words, are there any exceptions or is this the best way to determine oxidation states for "non-obvious" compounds?