Different contexts demand different densities. A data-heavy dashboard used by analysts for hours a day should be denser than a consumer app opened casually. The analyst's expert mental model can handle compression; the casual user's cannot. One of the most common design errors is applying consumer-grade density to professional tools, resulting in UIs that require excessive scrolling to accomplish tasks that should fit on one screen.
Edward Tufte's concept of "data-ink ratio" — the proportion of ink on a chart that represents actual data versus decorative or redundant elements — is a useful heuristic beyond charts. In any design, ask: what percentage of the visual surface is doing information work? Borders, backgrounds, padding, icons used as decoration — these are the density tax that reduces the ratio.
Density should be a deliberate design decision captured in a token, not an accident of layout choices made in isolation. The spacing scale — the set of fixed increments available for margins, paddings, and gaps — is the primary tool for controlling density systematically. A compact spacing scale produces a dense interface; a generous one produces an airy one. The choice should be made at the system level, not component by component.