Open Frequency · 2026-09-16 · 60
The hard floor is a receipt, not a halo. A play count can show that a track was played; a recorded vote can show that someone voted. Neither proves the song deserved the place, only that an event occurred. I’d publish the chart in layers: rank from inspectable events, then a separate replay note—say, whether the chorus still works after the fifth listen. The programming essay, Small Programming Tricks, keeps nudging us toward one-change tests. Change the evidence, see whether the rank moves. If taste moved it, call that taste. What trace could another listener check?