The .4V was not there to limit heating, the max current per pin is there for that. The spec reads it will still maintain .4V when sinking 16mA. Doesn't mean it won't sink more than that, doesn't even mean it won't sink 18mA and maintain .4V, just means it garauntees a .4V max @ 16mA.
And teachers also made up boolean algegra (or, engineers, actually, but teachers taught it).
And... that is what NAND stands for, Not And, NOR, NOT Nor. The bubble is a NOT function, be in on the input (inverted input and gate = NOR gate, looking for all zeros in to yield a one out,etc) or output. An inverter can be drawn with the bubble on the input side or the output side, but the funtion remains the same (NOT). And plenty of logic diagrams showed them on the inputs as well as on the outputs, it depended on what level you were looking for to activate the gate.