I'm not talking taxation is theft anymore. Im trying to figure out what sort of methodology you find acceptable for determining moral theories. Consistancy is part of a methodology. It may or may not be a good thing, but I have a hard time imagining a theory where consistency isnt important. Mabey you could explain why you arent sure about conistency.

This would just fall under punishment which is completely consistant. I'd agree that the amount of punishment is a grey area but I dont think its relative. If you agress against me, then I can agress against you.

The method I use, uses reality at it's core. The nature of reality reduces ethics to what actions are correct for a person to make that regard other people, themselves, and quite possibly other sentient beings. Consistency requires a sort of platonic forms that ethics must yield to. But in reality the types of topics that are difficult moral question don't have clear category distinctions, as the value of things appear to be highly subjective. It is what we believe to have value in that causes some action to be moral, correct, and some action to be incorrect, unethical. You appear to place great value in consistency, so you'd might be remiss to handle things in an inconsistent manner. I might value the lives of starving africans more than starving americans, or vise versa. It's this highly subjective value that we place on things that decide correct action from incorrect action not consistency. I just don't see what we gain by forcing our values to be perceived as consistent.

edit:
In addition, although the moral grammar that seems pervasive throughout cultures, history, and any other qualifier you could name appears to be generally similar. You know things like don't kill, suffering is bad, the golden rule etc. There is always some exception granted to these moral rules, moral rules that appear to have biological roots. The exceptions aren't applied consistently across really any metric. There doesn't appear to be any link between the types of exceptions that can be granted, there doesn't appear to be any link between the type of exception that are granted within a given population. These facts about reality lead me to believe that consistency isn't as important as you make it out to be.

The problem is that morality is very "intuitive" or "visceral", and so finding a rational, spelled out methodology may be quite contrary to how it works. It's why hypothetical scenarios like "save the child but kill 3 adults" are hard to answer.