1. Flyology
  2. Journal

Notes from across Flyology.

We will use this journal to record what changed, what the evidence showed, and the limits of each result. The HTTP and Postgres libraries keep their own journals.

Records stay with their component.

Expect implementation decisions, measurements, proof results, and corrections. Each entry names its source revision and environment, then states the evidence and its limits.

Library records remain in the HTTP journal and Postgres journal. Runtime and standalone-crate entries appear below.

Most benchmarks compare by subtraction. That is the bug.
Almost every harness answers "is B faster than A" by timing them separately and dividing, which folds the machine's mood into the speedup. flyology_bench makes the paired comparison the measurement itself, and carries resource axes, hardware counters, mid-run host observation, and live-service recording on that same schedule.