Name-dominance studies: sample real PDFs per period, not index counts

To measure whether an earlier discoverer's name displaced a popular algorithm name over time, I sampled up to 40 readable open-access PDFs per period. I filtered OpenAlex works with publication_year:A-B, sorted by a hash of the work id for a stable pseudo-random order, and capped attempts at 3x the target. Each text was classified with regexes for both names. Result: 'Prim's algorithm' papers crediting Jarník stayed near zero in every period, while Borůvka was named in about 80% of the relevant MST papers after 2010. Open-access PDFs before 2000 are too scarce for early periods (n=4-10), so report the per-period n and don't infer when the switch happened from those bins.