I'm not sure I can accept the way bluewing has even framed the question in the first place. There are definitely some metaphysical assumptions here that I'm not sure I like.
First of all, I'd like to attack the notion that there are "things" that are in and of themselves "good" or "bad." As first pointed out by AJ Ayer, to say that something is good or bad is not an empirical statement- there is no well defined test based on sensory experience we could carry out to see if such a statement holds. Thus, from a positivist perspective, ethical statements are meaningless. Statements such as "killing is right" and "charity is wrong" are not about killing and charity themselves so much as they are about our emotional reactions to killing and charity. To define ethics as the discipline that is concerned with what is right and what is wrong presupposes the existence of well-defined categories "right" and "wrong" which exist a priori- our job is just to figure out how to put the world in these categories. This is getting awfully close to presupposing platonic metaphysics to even get ethics started in the first place. The existence of such categories is not an empirical question. I would assert that from a positivist perspective, ethical claims are not true or false.
I would also like to contest the claim that the ultimate goal of all human activity is the acquisition of happiness. From an evolutionary perspective, the animal that maximizes its own chance of survival and reproduction is the one that is most likely to pass on its genes, not the animal that is the most happy. Thus, it follows that our behaviors and emotions are designed in a manner that will lead to action that preserves the survival of our species. Achievement of long term happiness is not conducive to maximizing survival chances if a strong dose of fear and self-hatred is useful in getting us to go out into the wild and go hunting. Furthermore, the structure of our brain, designed by evolutionary forces, is not designed to fit into a simple philosophical model of human behavior. As argued in "Stumbling on Happiness" by Daniel Gilbert, we simply do not know what makes us happy much of the time. If we are to take bluewing's ethical system then, we simply do not have any way of learning about ethics, because we are unreliable judges of our own future happiness.-
I'd also like to propose an alternative formulation of ethics, based upon Wittgenstein's assertion that ethics and aesthetics are one and the same. The foundation of ethics is the alogical (but most certainly NOT illogical) "value judgement" that something is ethical or unethical. To say that something is ethical is a statement about your perception of the aesthetic value of the world, not a statement about the object itself. Killing is wrong because there is something emotionally awful about the destruction of another life, and the only foundation of such a judgement is our emotional response, which is independent of logic and reason. This is entirely analogous to how we judge a work of art, whether a picture of a puppy is adorable, or anything else that is not solely empirical.
This alogical foundation of our ethics is not exactly the subject matter of studying ethics. While fundamentally it is this emotional foundation that is really what matters, we cannot speak of it directly. This is where human empathy becomes useful. Whether philosophical discussion, art and literature, or less "highbrow" (but not less valuable) methods are at hand, the purpose of ethics is to extract the commonalities of our ethical/aesthetic reactions to the world. Language and culture are organic entities based upon the signaling games we play to coordinate our inner emotional states. We cannot talk about these states directly, but human behavior and its empathic foundation allow us some power in communicating them, even if they must exist outside of the method of their communication. I would define an action to be ethical if that is simply the consensus value judgment expressed in a culture. "In a culture" is needed so that these methods of communication can actually coordinate people's emotional states, since there is no good notion of consensus of people who have no method of agreeing or disagreeing.