The ... zone system was ... designed to get a wide range of tones captured onto film ....
Thanks for making that point, along with some points that I did include in quoting your post.
That points to the topic of how to best specify exposure and film speed.
I am familiar with the concept that, historically speaking, our current iso system has its roots in research done at Kodak where they exposed a series of negatives, made the best print possible from each negative, and then asked a panel of people which prints were excellent. The film speed scale was based on what minimum exposure was required to produce an excellent print. There have been some simplifications/refinements since then, but that is the gist of it.
When it comes to capturing a wide range of tones onto film, I think the "first excellent print" method probably missed the point if the goal is to capture the widest range of tones. Admittedly, capturing the widest range of tones was not the goal of the research that ultimately lead to our current ISO film speed scale.
Here are my thoughts about capturing the widest range of tones. To capture the widest range of tones (rather than determining the minimum exposure that can produce an excellent print) the best approach would be to go back to basing the system on gamma rather than on contrast index or related quantities such as G-bar. If we make the simplifying assumption that a typical characteristic curve is approximately a symmetrical sigmoid, then basing the exposure on the middle of the straight line portion of the characteristic curve will give the widest tonal range captured on film. This implies that the system would be based on the slope of the relatively straight portion of the characteristic curve, i.e. based on gamma. (Sometimes people incorrectly equate gamma with contrast index or G-bar, but those quantities are not the same either by definition or by typical numerical value.)
There are some mathematical reasons for making that argument, but let me try to explain with a mix of mathematical and qualitative comments.
One idea behind my suggestion is that one should usually place the subject of interest so that it's average density lies on the part of the curve that has the highest contrast. (Note: this also means that one meters off from that point.) That point is the center of the relatively straight part of the characteristic curve, and the slope of the curve is at its maximum at that point. Mathematically speaking, that point is defined by the second derivative of the curve (i.e. a value of zero for the second derivative), and the contrast is determined by the slope of the curve at that point, i.e. by gamma (i.e. the first derivative of the curve). That will give you the maximum tonal separation for the subject of interest. The darker and lighter portions of the photograph will then fall where they may, but since they are not of primary interest we accept the compromise that their tonal range will not be as well separated as the tonal range of the subject of interest. That compromise is pretty much inevitable no matter what system one uses. I am just stating it explicitly here for sake of clarity.
Here is an example of how this would work. This figure is from some actual film/developer testing I did.
In this case, if I could set my meter at an iso of 54 it would give the same density that the graph is showing at the gamma point (density = 1.145, gamma = 0.820). Therefore, I would classify this film as having an exposure index of 50.
The lower point on the graph marked as slope = gamma/3 has a density of 0.516, which as luck would have it is very close to the ideal speed point of 0.525 (base plus fog was 0.425), and that point is 3.5 exposure stops below the gamma point. The upper point with a slope of gamma/3 has a density of 1.870 and is 4.3 exposure stops above the gamma point. Thus, in this case, the curve is not quite symmetrical, but that detail is not too important from a practical point of view.
Interestingly, the parameters for this curve are very close to the iso definition (average slope of 0.64 for my curve vs. slope of 0.62 at a density of 0.8 units above the iso-defined speed point), and in this case I also get a predicted exposure index close to half way between 50 and 64, which is almost the same as what I get from my proposed method.
One could say, "since they give essentially the same result, why use your proposed method instead of the iso method?" That would be a good point. However, I think that if one is interested in recording the widest tonal range on film then my conceptual starting point and method is more defensible than the iso method, which takes as its conceptual starting point to achieve an adequate level of shadow detail. I guess one could summarize the philosophical difference as saying that my method emphasizes the main subject whereas the iso method emphasizes the shadows. In this particular example they end up at about the same end point, even though their conceptual starting points were different. For other film/developer combinations the two approaches might not give a similar recommended exposure index.
Please note that one advantage of my proposed method is that if two film/developer combinations are set up for the same gamma then, regardless of whether or not the same exposure indexes are recommended for the two film/developer combinations, if the recommended exposure indexes for the two film/developer combinations are followed they will give essentially the same tonal separation for the main subject of the photograph.
By the way, I would call this particular film/developer combination as long-toe. What do you think?
One last point: If one wants to emphasize the shadows a bit more one could shift the effective film speed down a little bit without greatly affecting the tonal separation at of the main subject, provided the one doesn't go too far in that direction. Conversely, one could shift the effective film speed up a little if one wants to give more tonal separation to the highlights without greatly affecting the tonal range of the main subject, provided one doesn't go too far in that direction.