I think this is because it is more nuanced that you, or most people, are giving credit to. Privacy is not a binary option of private vs public but rather a continuum. This is where "reasonable expectation" falls and tries to create a division, a third option. But I think we need more since the environment has changed. As an example, if in a public place you pull yourself and your friend to a dark corner that's obscured you have increased your expectation of privacy. Because you're literally "hiding." The extreme of this might be a public bathroom where you definitely do have a legal expectation of privacy despite being a public place.
So the problem here is that we're trying to make clear cut discriminators when the world has gotten a lot messier. Unfortunately legal systems require discretization, because it is easier to legislate clear boundaries (laws are piecewise functions -- if-then-else). But the world is messy and it needs to be recognized that those discrete boundaries are fuzzy too (i.e. they are guides, not hard fixed and objective boundaries). Probably also coupled with human propensity to discretize continuous variables due to the compression effects. This really is why we have many "know it when I see it" definitions rather than hard rules and you may notice that as complexity is increasing (making this more common) and so is the disagreement with said loose definitions. Then we have a core problem of communication breaking down because everyone is working off a different set of assumptions, assuming everyone else has the same assumptions, and unwilling to accept other bases because our assumptions are objectively generated ;)
So the problem here is that we're trying to make clear cut discriminators when the world has gotten a lot messier. Unfortunately legal systems require discretization, because it is easier to legislate clear boundaries (laws are piecewise functions -- if-then-else). But the world is messy and it needs to be recognized that those discrete boundaries are fuzzy too (i.e. they are guides, not hard fixed and objective boundaries). Probably also coupled with human propensity to discretize continuous variables due to the compression effects. This really is why we have many "know it when I see it" definitions rather than hard rules and you may notice that as complexity is increasing (making this more common) and so is the disagreement with said loose definitions. Then we have a core problem of communication breaking down because everyone is working off a different set of assumptions, assuming everyone else has the same assumptions, and unwilling to accept other bases because our assumptions are objectively generated ;)