It's the idea of "counting" grains which I was objecting to, as they overlap both on the horizontal plane and on the "depth" of the film and so what we see is not a "number" of grains composing an image (like in a mosaic, or in a pointilliste painting, or in a digital picture) but the overlapping of borders of these grains, which makes the "count" of them misleading.
This very moment my screen shows me characters which are 1 pixel wide. I can clearly see, in text, the width of the pixels. They are "ordered" in a matrix and so counting them makes sense to asses the "resolution"* of the screen.
The "resolution" of the digital image (or the mosaic, the TV screen etc.) is strongly correlated to the number of pixels. Even if we don't perceive the single pixels (we certainly don't while looking at a move on the TV set) we perceive that the resolution depends on the "count" of the pixels. High definition TV gives perceivably a higher definition even if we cannot isolate the single pixels, we perceive there is a higher count of them forming the image, because the image is formed on one single plane of geometrically ordered tesserae.
This analogy doesn't stand well with "grains" because they behave differently. In colour film the clouds of filaments are "transparent" and their overlapping gives various areas of colours where the variation in colour can be continuous at a certain magnifying scale (well shown in the Vitale document) so that one doesn't have a matrix of pixels with an abrupt variation of colour from an atomic information to the next, like in digital, or in a mosaic.
Digital and analogue are IMO a bit like mosaic and watercolour. In a mosaic we can count the number of tesserae but in a watercolour we cannot "count" an equivalent, that's what I meant. Probably somewhere at some microscopic scale we can find some solid "grains" on the paper but that is on a completely different scale, and visual effect, if compared to the mosaic.
With black and white probably the filaments are more "opaque" and, although the darker or lighter effect is always given by the overlapping of crystals edges in the "depth" of the film (in colour film a yellow "cloud" can partially overlap with a magenta cloud and partially overlap with another magenta cloud, how do we count them, in black and white we have a section where the "dirtier" the water the "darker" the image, so to speak, and we could theoretically count the single particles of dirt in suspension, each being, I guess, quasi non-transparent), one can say that the "working" is a bit more similar to the digital medium because there aren't, IIRC, the "half transparent" crystals which happen to be in colour film (the filaments forming the clouds).
* OK "resolution" itself can have several meanings.
I don't think we disagree. It depends on which microscope scale we consider "meaningful" for the comparison with digital.
Yes, I think we just read the OP's question in different light. I was getting it in the sense of a bit humorous or "child-like" search for trivia, such as "how many grains of sand in Sahara desert?" -- So it actually wouldn't make much sense but would count as a "fun fact". Especially, when said to a digital person counting his megapixels :).
A possible complication lies in the fact that the graininess of the final picture, a positive, is actually a map of the spaces between the grains of the negative.
Many times when I have looked at a very dense part of a negative through a grain magnifier on the baseboard of an enlarger I've remarked on the smoothness of the view. It's mostly dark with bright points here and there but I know those few widely separated "stars" will deliver conspicuous pepper 'n salt grain in the highlights of my final photograph.