User Tools

Site Tools


provenance:privacy:requests

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revisionPrevious revision
provenance:privacy:requests [2026/09/17 07:24] – Markup rendering sweep: wrap escapes, CLI flag typography, and heading quotes. Authored by Claude. karel.kubicek.claudeprovenance:privacy:requests [2026/09/21 14:57] (current) – Amendment 2026-09-21: record the removal of the duplicate which-lists table, the stale 198 pointer (filter_lists prints 194), the script output behind every figure kept, and the rendered-DOM structural check. Authored by Claude karel.kubicek.claude
Line 1020: Line 1020:
  
 Mechanical rendering repair only: a fresh live raw/XHTML export of 188 pages was checked with ''check_wrap.mjs'' and ''check_typography.mjs''. Affected plugin tags, CLI flags and heading markup were repaired; no figures or substantive prose were changed. The resulting source and rendered DOM were re-checked after saving. Mechanical rendering repair only: a fresh live raw/XHTML export of 188 pages was checked with ''check_wrap.mjs'' and ''check_typography.mjs''. Affected plugin tags, CLI flags and heading markup were repaired; no figures or substantive prose were changed. The resulting source and rendered DOM were re-checked after saving.
 +
 +===== Amendment, 2026-09-21: the duplicate "which lists" table was removed =====
 +
 +**What was wrong.** This page's //Use in Publications// carried a full per-list adoption table over its S1 population (197 papers), and [[programming:filter_lists]] carried the same table, computed by the same fold, over a 194-paper population. Two copies of one table, maintained separately, drift; [[provenance:programming:filter_lists]] recorded the duplication as a **live inconsistency left for a later run to finish**, and this is that run. A second, smaller defect came with it: the sentence introducing the table described the filter-lists population as **198 papers**, which was that page's figure before its tuple-context exclusion. That page has printed **194** since, so the pointer was quoting a number the target page no longer published.
 +
 +**The commands and their real output.**
 +
 +<code>
 +$ node scripts/report_requests.mjs
 +S1 — used or produced an ad-or-tracker filter list                197
 +S2 — classified web requests for an advertising/tracking purpose  164
 +**page population = S1 ∪ S2**                                     254
 +
 +Denominator: 197 papers that used or produced an ad-or-tracker filter list.
 +EasyList                                                        110     55.8%         32
 +EasyPrivacy                                                     71      36.0%         26
 +Disconnect                                                      48      24.4%         28
 +Ghostery / WhoTracks.me                                         34      17.3%         12
 +── Exact-string vs folded, for the four lists the page names ──
 +  EasyList                 exact-string  90  folded 110  undercount 18.2%
 +  Disconnect               exact-string  26  folded  48  undercount 45.8%
 +
 +$ node scripts/report_filter_lists.mjs
 +   USED or PRODUCED a filter list  == POPULATION               194 3.3% of corpus
 +</code>
 +
 +**What changed.** The fifteen-row table and the //Folding is not cosmetic here// paragraph were replaced by two paragraphs. The first keeps the four figures a reader of //this// page needs — 197, and EasyList 110 (55.8%), EasyPrivacy 71 (36.0%), Disconnect 48 (24.4%), Ghostery / WhoTracks.me 34 (17.3%) — together with the folding lesson stated at its strongest (90 against 110, and 26 against 48), and points at [[programming:filter_lists#Which lists|the canonical table]]. The second explains the 197-against-194 gap as a property of two membership rules rather than a disagreement. **Neither count was changed and neither was reconciled**, which is the standing instruction for this pair: both scripts import ''req_fold.mjs'', so the fold is identical and only the membership rule differs. The engines-and-parsers sentence was kept — it is not duplicated on the filter-lists page — and the S1/S2 narrative above it is untouched.
 +
 +**Rendered-DOM check.** Tables on the page went **14 → 13** and rows **114 → 98** (one header plus fifteen list rows). Headings (28), list items (60) and ''<pre>'' blocks (0) are unchanged, and the page carries **0** ''wikilink2'' red links. The ''%%[[programming:filter_lists#Which lists]]%%'' pointer resolves against the real heading id ''which_lists'' on the target page.
 +
 +**Reviewers.** One ''sonnet'' figures-against-script pass, covering the edited passage and the sections either side of it. No citations pass: no ''{[citekey]}'' and no quoted claim was touched.
  
provenance/privacy/requests.1789629843.txt.gz · Last modified: by karel.kubicek.claude