wiltw, to clarify what Mark means regarding speed point, I suggest you read Kodak document H-740, "Basic Photographic Sensitometry Workbook", which explain the basics of sensitometry, and I certainly won't go a centimetre past that document as far as complexity is concerned
What Mark means is a well-known property of B&W negatives: by altering the developing time you alter much more the highlights than the shadows. One can imagine the characteristic curve as "pivoting" around a shadow point, on the left of the graph, where there is minimum density. The more the exposure, the higher the density, by a "law" which is defined by the ISO speed of the film but also by the development time (or agitation, or temperature). The curve can "pivot" around that shadow point at different kind of development and while the different development raises somehow all the curve,
the effect will be more relevant to the highlights.
That leads to the "zone system" by Adams. By developing negatives one by one, you can alter the contrast and the highlight renditions of the image. That's because the development can alter the highlights much more than the shadows. So you control the shadows with the exposure, and you "control" the highlights with the development. In this manner you can obtain a negative which is more contrasted, or less contrasted, than the original scene, and always print on the same gradation of paper with the intended result.
This is all very esoteric, very sophisticated photography. The Zone System applies only in Black & White, in sheet film (because you develop each shot individually), and is as far as I know quite a complication in a world of variable contrast papers. But that's only what I heard through the grapevine, I use slides, and I would like to go back to it.
[My example of the digital ETTR was intended to mean that it is a technique that uses a "non ISO" exposure to reach a certain result. It's not meant to be taken literally and transported in the analogue world].
Before we talk slides, we should talk very simple sensitometry. For a complete ignorant of the matter like yours truly, you can define a speed scale either by choosing a certain arbitrary density on film, and then measuring how much exposure it takes to arrive to that density (the higher the exposure, the lower the speed) OR by choosing a certain arbitrary exposure on film, and then measuring how much density that produces on film (the higher the density, the higher the speed).
There certainly are very many possible algorithm for defining film speed. The most famous are ASA/ISO, and DIN. There used to be others, created in the United Kingdom, in Eastern Europe, in Russia. Now the world set on the ISO standard which is basically the ASA standard.
If I were Mr. ISO I would define a certain density on film "around the middle" and measure how much exposure it takes to arrive there. Mr. ASA did things very differently. He chose a certain point at the near transparent end of the characteristic curve of the negative, and from the density of that point, he derived a certain speed, assuming, I suppose, that the rest of the curve was basically linear and the ASA number would describe the "entire curve" (or line) of the film even if it was calculating by extrapolating the very initial part.
As a completely ignorant person of the subject, I can infer that Mr ASA chose to measure density near the extreme left because that minimizes the differences in development. Even though the ISO standard certainly stipulates a certain "standard" treatment defined by the manufacturer, measuring on the left of the curve takes some complexity out of the tests, takes away some head-scratching.
Conceptually, though, given a standard well-known development, Mr. ASA could have chosen any density, and is possible that Mr. DIN did something completely different than Mr. ASA, considering that certain films (Vericolor III if memory serves) have a "different" speed rating in the ASA and DIN systems (meaning: they employ two different algorithms). (The difference was 1/3 of a stop but is relevant for what I am saying). And the aim of Mr ASA, just as the aim for Mr. DIN, is to actually show the behaviour of the film around the middle part of the curve. If Mr. ASA chooses to take a certain low density point as "speed point" he does it for some technical reason but not because he has Zone System in mind, so to speak! Mr. ASA does not think that the user will develop for the highlights, or use sheet film. He means to tell you the speed of the film all along its linear part of the curve so that Aunt Sue will get the holiday pictures with the same tones that she saw during her trip to that most beautiful wonder which is Rome (shameless plug).
Now, back to slides. Slides matter because all this talk about exposing for the shadows, developing for the highlights, compensating in print does not stick with slides and does not stick with aunt Sue. I insist: slides can be developed only in one correct way, and can be seen only in one way, there's no print and there's no "recover" of details, no dodging, no burning, the bunny must come out of the hat at first attempt.
With slides, a light meter must work exactly and very precisely as intended. And it does! So - I insist - there is, even "implicitly" in the ISO speed determination, a way to make the middle grey (and all surrounding tones of all colours) fall exactly where they are perceptually faithful. Yes the extremes of the dynamic range will not be rendered, or will be rendered with some faults. But the tones around the average MUST be perceptually precise on slides, and they are. So lightmeters are actually devices that allow us to know where the
middle tone will fall, and they show that to us every day. And they must be calibrated for a certain, unknown but certainly determined by the manufacturer, shade of middle tone. And I would be very, very surprised if this middle tone wouldn't fall exactly where the human eye sees middle tone, and where the "world at large", as shown by a research conducted by Kodak and mentioned by Ralph, shows middle tone. That's 18%.
So let's keep the reasoning alive because there is a lot to be understood more precisely (I have got a couple of huge doubts that I will try to clarify with your help) by keeping slide film in mind. Let's forget B&W negatives for the time being.