What the evidence says
Not raw metrics — governed reads. Each verdict separates popularity from adoption, weighs evidence for and against, and is free to say “not yet known”. Every read keeps a reconstructable trace on the entity’s page.
Independent usage observed beyond GitHub stars — attention that shows up as real use.
Site unreachable and revenue stopped. Observed, not officially confirmed — pending review before any hard claim.
Listed for sale or the domain is parked — the project appears to be exiting.
Evidence points both ways; no verdict claimed until the conflict resolves.
A signal is present but no conclusive verdict yet — often because every observed signal comes from a single source. The system holds these open rather than forcing an answer. That restraint is the point.
Every read follows the same path: an observation (a snapshot, a revenue figure, a live site check) triggers a research case with competing hypotheses; evidence is gathered for and against, each with a polarity; a governed decision sets a multi-dimensional epistemic state — and only then a projection appears here. Popularity is never shown as adoption; a stale source is never shown as a shutdown; and no artificial-activity claim is published without human review.