Old Revisions

These are the older revisons of the current document. To revert to an old revision, select it from below, click Edit this page and save it.

2026/08/27 12:39 programming:interaction – Reviewer 4 (generic): add 'Interacting is not free' — click fraud, form side effects, load multiplied by 3x, and the visibility trade-off — and a Monday-morning default design; replace the unaudited 'none of them justified anywhere' with the measured just karel.kubicek.claude +5.5 KB (current)
2026/08/27 12:34 programming:interaction – Self-review: replace an unmeasured 'almost no methods section contains' with a claim that can be checked (no reporting checklist asks for it). Authored by Claude karel.kubicek.claude +34 B
2026/08/27 12:34 programming:interaction – Self-review: 'everything on the checklist is in there' overstated — the Aziz and Wilson quote covers five of the seven items; name the two it does not and where they live. Authored by Claude karel.kubicek.claude +336 B
2026/08/27 12:32 programming:interaction – Self-review: scope the open-questions box's 'nobody' claims to the corpus's seven venues, and the 'one paper that did measure' likewise. Authored by Claude karel.kubicek.claude +215 B
2026/08/27 12:31 programming:interaction – Self-review: two 'no paper mentions bot detection' claims came from a first-person-sentence probe, and a probe's zero is not 'nobody' — three of the 80 papers in that bucket do mention it. State the probe's scope and the loose count. Authored by Claude karel.kubicek.claude +229 B
2026/08/27 12:31 programming:interaction – Self-review: 'median 10, unchanged across all five buckets' was asserted, not computed — the per-bucket medians are 5, 5, 10, 15, 5 on 8/11/37/33/15 papers. State the real series and call it small-sample noise rather than a trend. Authored by Claude karel.kubicek.claude +253 B
2026/08/27 12:30 programming:interaction – Self-review: 'almost nothing on a modern page loads at load' overstated; name what actually does not (lazy ad slots, tabbed content, form-triggered behaviour). Authored by Claude karel.kubicek.claude +133 B
2026/08/27 12:29 programming:interaction – Quantify the shared-crawlConfig-quote blindness rather than asserting it: on a reproducible 12-paper sample, 6 of 12 shared quotes evidence the depth, 5 evidence a different key, 1 is about visit structure. Authored by Claude karel.kubicek.claude +270 B
2026/08/27 12:28 programming:interaction – Add 'What a Complete Statement Looks Like': the two sentences from Aziz and Wilson that answer every item on the reporting checklist, including the one detail most often left implicit (subpages get the same treatment as the landing page). Authored by Clau karel.kubicek.claude +974 B
2026/08/27 12:27 programming:interaction – Self-review: 'clicking is flat since 2010' overstated a series whose highest bucket is the earliest (43.8% in 2010-2013, 29.9-35.3% after); state the range instead. Same for 'rounding errors' -> 'under 4%'. Authored by Claude karel.kubicek.claude +75 B
2026/08/27 12:25 programming:interaction – Self-review: 'by a wide margin, the best-reported crawl-configuration field' was overstated — the margin over authentication is 5.5pp. Replace the claim with the measured ranking of all six crawlConfig fields, which makes the real point better (depth 75.1 karel.kubicek.claude +481 B
2026/08/27 12:22 programming:interaction – Reviewer 2 (citations/quotes): replace the Musch and Johns paraphrase-in-quotation-marks with the paper's own sentence, and add its real denominator (206,676 sites with an indicator, 30,494 subpage-only) plus the per-technique spread the +17% aggregate hi karel.kubicek.claude +583 B
2026/08/27 12:17 programming:interaction – Reviewer 1 (figures vs script): drop the SSO count, which came from the 1,080-paper crawlConfig population and is zero in this page's 857; republish both fixture files byte-identical to the committed scripts; make the probe print the Chromium build and pl karel.kubicek.claude +1.1 KB
2026/08/27 12:09 programming:interaction – Replace the 2.4 KB stub with a full page on crawl interaction: the four axes (depth, subpage selection, on-page action, identity), corpus figures with denominators (857 web crawls, 417 on the depth axis), the Urban/Aqeel disagreement, the 100-subpage satu karel.kubicek.claude +43 KB
2026/08/21 14:50 programming:interaction – Boxes: <wrap> renders a span, use uppercase <WRAP>; one box per list, not per bullet karel.kubicek.claude ±0 B
2025/03/19 08:51 programming:interaction – Links and references karelkubicek +349 B
2025/01/15 09:09 programming:interaction – created - brainstorming admin +2 KB