Originally posted by Class A AFAIC, there is no such thing as "accurate exposure" that can be mechanically determined. Sometimes you want skin darker, sometimes you want it brighter; the light meter will always just give you one interpretation and I don't see why making sure you captured the full tonal range by using a histogram is any worse.
It really depends on one's definition of "accurate exposure". I use two phrases when talking about exposure accuracy, "proper exposure" and "good tonality", both of which are important, but I use them to mean distinctly different things. By my definition, "proper exposure" is achieved when the subject renders exactly as he/she/it did in real life (The scene is less important, but that's the topic for another post). Light stuff will render light, dark stuff will render dark, and middle gray will render exactly middle gray. That's what an incident light meter tells me, 'how much light is falling on this subject?' and I normally base my exposure on that. What you get ideally is a print that could be held next to the original scene with no discernible difference of luminance values. That works well for portraits, but not as well for landscapes. Of course, it goes without saying that one needs to use a color managed workflow for the results to have any meaning.
"Good tonality" on the other hand deals with controlling the contrast of a scene from capture to print which may be wider or narrower than that of the final medium. A picture exhibits "good tonality" (usually) when there is a full range of tones from black to white and good retention of detail in both the highlights and shadows. This is what Ansel was after, and this is what the histogram tells you. The histogram is, in a nutshell, the zone system distilled. A scene with a wider tonal range than the paper can reproduce must be "scaled down" to fit within the limitations of the medium, and objects, particularly at the far ends of the tonal scale, can rarely be "properly exposed" if the image is to look realistic. Even objects towards the center of the tonal scale can experience slight shifts in tonality when mapping a particularity contrasty scene like a sunset to a piece of paper, or even a computer screen.
"Good tonality" is tremendously important in black and white photography, probably more so than "proper" exposure, but in color photography, proper exposure should not be ignored. This is because exposure is inextricably linked to saturation. An overexposued picture tends to look washed out, this is because colors like blue and red cannot reach their highest levels of saturation (chroma) at higher luminance levels. They are, essentially, darker colors. Conversely yellows will tend to look muddy when underexposed. Therefore "proper exposure", which is to say exposure based on the amount of light falling on the subject is the only way to get truly accurate colors across the board.
Now, can one determine "proper exposure" through reflective readings? Sure, and 5 dollar 18% gray card goes a long, long way towards achieving the accuracy of an incident light meter, but it's a pain in the butt trying to judge manual flash/strobes that way. I use my light meter religiously, in fact, I rarely use my camera's meter for anything. I feel like I'm at the mercy of the machine when I have to switch to Av or Tv mode, consequently I only do it when there is no time to use my hand held meter, and luckily, as a portrait shooter, those times are rare. But that's just how I shoot, not that my way is any more or less correct than anyone else.