This paper introduces a practical, repeatable method for reading Zenodo search results as observational evidence: to distinguish settled vs unsettled terminology, detect concepts being re-invented, locate unresolved framing problems, and identify where attention gravity is forming across technical domains. The method is designed for scientists and engineers who already read dense material and want higher signal: a four-pass scan (Best Match → Newest → Most Viewed → Journal/Newest) that exposes unstable language, recurring claims, and latent demand — not popularity. The goal is not to “win search.” The goal is to produce a disciplined way to find what should be written next, and why.
David Forbes (Sun,) studied this question.