Perceived densities?

In the current issue of Access, Eric Eidlin writes about "What Density Doesn't Tell Us About Sprawl."

Urban planners love to talk about density and sprawl. Trouble is that both are hard to define. Eidlin cites recent work on "perceived" density. But this is really a weighted average of densities across a metro area's small spatial units (he likes census tracts).

I have previously blogged about problems with the density idea. Creative people are apparently comfortable in low-density Silicon Valley as well as in high-density Manhattan. How can that be? It is the whole package that matters. I am not sure that the whole package can be assessed via the cited "perceived" density measure, but if analysts insist on looking for simple ways to characterize whole metropolitan areas, then let them. In this case, I worry that calculating the weighted averages of Silicon Valley vs Manhattan densities will not provide much in the way of new insight.