The grain size of the salt is not taken into account. In this sense, the model takes a telescopic view of stars–that they are points of light. A model that did include the vast differences in stellar diameters would quickly loose its usefulness as a simplified but graspable metaphor, in part because we’d need a kind of salt with greatly variable grain size that also happened to account for the distribution of different stellar sizes in the galaxy. I know we have a sense of where the Milky Way stars fall on this distribution, but you could see where it would quickly get out of hand as a teaching model.

For distances, it’s the same issue; accuracy makes the model untenable. Obviously stars are not distributed across the galaxy equally, with far greater numbers in and near the core. That we know, but we’re still learning what the rest looks like. The model of the many-armed Milky Way that I grew up with is now quite different, and likely to change again. The model runs into the limits of our present knowledge. We’re not even sure about how many stars there are in the Milky Way–and the discrepancies in estimates are not small (+- 100 billion!). So the 7 miles is meant to represent the average density, but I’m not sure it’s even as representative as that. The point of the 7 miles it that it’s far–very, very far, between most stars, and to give people a basic building block with which to start to construct the bigger picture. Which of course quickly grows past our comprehension.

I’m not entirely happy with those abstractions…I wish I could find a more defensible number than 7 miles. I’ve looked and can’t seem to find much on average interstellar distances. I’d love to refine the model further, however.