[brundolf]

This was basically it. The dataset was the same across users so caching was simple and efficient, and the front-end had no difficulty handling this much data (and paging client-side was snappier than requesting anew each time)