SEMANTIC HTML · ARCHIVED RESULTS

Semantic HTML.
Compare what tools preserve.

The same text. One markup change.
Two outputs to inspect, with evidence.

Explore an archived run from 20 September 2026, re-evaluated on 21 September. No extraction runs during your visit.

EXPLORE THE CONTROLLED CORPUS

Table

newspaper4k 0.9.3.1 · Extractor / reading tool · French document

All 42 comparisons

Values must match exactly in the same row. An XML caption is not a column header; a Markdown table does not prove the input’s HTML semantics. Test scope.

A

VARIANT A

Headers with th + scope

Output produced

Expected value row : Exact row not found · Column headers : Relationship not demonstrated · Row header : Relationship not demonstrated. Source

01 · The HTML change

<table><caption>Mesures synthétiques</caption><thead><tr><th class="case-cell" scope="col">Indicateur</th><th class="case-cell" scope="col">Unité</th><th class="case-cell" scope="col">Période</th><th class="case-cell" scope="col">Valeur</th></tr></thead><tbody><tr><th class="case-cell" scope="row">Temps de réponse</th><td class="case-cell">secondes</td><td class="case-cell">mai 2026</td><td class="case-cell">1,8</td></tr></tbody></table>

Exact input excerpt · Original HTML ↓

02 · What the tool produces HTML

<div><h1>Protocole hors ligne de conservation des relations</h1><p>Ce document synth&#233;tique sert &#224; comparer deux balisages &#224; texte, ordre, classes et environnement identiques. Les donn&#233;es sont fictives et ne d&#233;crivent ni Edikka ni un client.</p><h2>Mesures synth&#233;tiques</h2>Mesures synth&#233;tiquesIndicateurUnit&#233;P&#233;riodeValeurTemps de r&#233;ponsesecondesmai 20261,8<p>La conclusion doit &#234;tre relue avec son unit&#233;, sa p&#233;riode, son p&#233;rim&#232;tre et sa r&#233;serve. La pr&#233;sence d&#8217;un mot dans une sortie ne suffit pas &#224; prouver que la relation a &#233;t&#233; conserv&#233;e.
…

Excerpt from the start of the output; it does not establish absence on its own. Full output below.

03 · The documented observation

Expected value row
Exact row not found
Column headers
Relationship not demonstrated
Row header
Relationship not demonstrated
Output structure
Absent

Verdict source: results.json · table-fr-1 / newspaper4k

Inspect the full evidence

The verdicts above come from corrected results.json. The raw file’s evaluation fields are historical and are not authoritative here.

results.json/cases/2/evaluation/relation/newspaper4k

Corrected verdict, unchanged
{
  "status": "ok",
  "tool_class": "content_extractor",
  "format": "html",
  "table_structure": "absent",
  "expected_data_row": "not_found",
  "column_header_relation": "not_explicit",
  "row_header_relation": "not_explicit",
  "column_labels": [],
  "data_values": [],
  "evidence": {
    "header_row": null,
    "data_row": null,
    "table_excerpt": null
  },
  "output_field": "additional_extractors.newspaper4k.html",
  "output_sha256": "83fe1f59de6678c68e8fe843a9fa601f606870347da6a587963c95e8869902be",
  "source_column_headers_explicit": true
}
Complete studied output field

additional_extractors.newspaper4k.html

<div><h1>Protocole hors ligne de conservation des relations</h1><p>Ce document synth&#233;tique sert &#224; comparer deux balisages &#224; texte, ordre, classes et environnement identiques. Les donn&#233;es sont fictives et ne d&#233;crivent ni Edikka ni un client.</p><h2>Mesures synth&#233;tiques</h2>Mesures synth&#233;tiquesIndicateurUnit&#233;P&#233;riodeValeurTemps de r&#233;ponsesecondesmai 20261,8<p>La conclusion doit &#234;tre relue avec son unit&#233;, sa p&#233;riode, son p&#233;rim&#232;tre et sa r&#233;serve. La pr&#233;sence d&#8217;un mot dans une sortie ne suffit pas &#224; prouver que la relation a &#233;t&#233; conserv&#233;e.</p></div>
Archived version and options

newspaper4k 0.9.3.1 · Execution 2026-09-20

{
  "version": "0.9.3.1",
  "configuration": "No configuration field in this output; consult the archived runner."
}

Archived runner · Python options

SHA-256 hashes
HTML input
7e19912f8f4337ceedbdb6b7d9cc0b43aebf2ed8f72d359f411e9a725142c556
Evaluated output field
83fe1f59de6678c68e8fe843a9fa601f606870347da6a587963c95e8869902be

Complete raw output file ↗

B

VARIANT B

Ordinary td cells

Output produced

Expected value row : Exact row not found · Column headers : Relationship not demonstrated · Row header : Relationship not demonstrated. Source

01 · The HTML change

<table><caption>Mesures synthétiques</caption><thead><tr><td class="case-cell">Indicateur</td><td class="case-cell">Unité</td><td class="case-cell">Période</td><td class="case-cell">Valeur</td></tr></thead><tbody><tr><td class="case-cell">Temps de réponse</td><td class="case-cell">secondes</td><td class="case-cell">mai 2026</td><td class="case-cell">1,8</td></tr></tbody></table>

Exact input excerpt · Original HTML ↓

02 · What the tool produces HTML

<div><h1>Protocole hors ligne de conservation des relations</h1><p>Ce document synth&#233;tique sert &#224; comparer deux balisages &#224; texte, ordre, classes et environnement identiques. Les donn&#233;es sont fictives et ne d&#233;crivent ni Edikka ni un client.</p><h2>Mesures synth&#233;tiques</h2>Mesures synth&#233;tiquesIndicateurUnit&#233;P&#233;riodeValeurTemps de r&#233;ponsesecondesmai 20261,8<p>La conclusion doit &#234;tre relue avec son unit&#233;, sa p&#233;riode, son p&#233;rim&#232;tre et sa r&#233;serve. La pr&#233;sence d&#8217;un mot dans une sortie ne suffit pas &#224; prouver que la relation a &#233;t&#233; conserv&#233;e.
…

Excerpt from the start of the output; it does not establish absence on its own. Full output below.

03 · The documented observation

Expected value row
Exact row not found
Column headers
Relationship not demonstrated
Row header
Relationship not demonstrated
Output structure
Absent

Verdict source: results.json · table-fr-2 / newspaper4k

Inspect the full evidence

The verdicts above come from corrected results.json. The raw file’s evaluation fields are historical and are not authoritative here.

results.json/cases/3/evaluation/relation/newspaper4k

Corrected verdict, unchanged
{
  "status": "ok",
  "tool_class": "content_extractor",
  "format": "html",
  "table_structure": "absent",
  "expected_data_row": "not_found",
  "column_header_relation": "not_explicit",
  "row_header_relation": "not_explicit",
  "column_labels": [],
  "data_values": [],
  "evidence": {
    "header_row": null,
    "data_row": null,
    "table_excerpt": null
  },
  "output_field": "additional_extractors.newspaper4k.html",
  "output_sha256": "83fe1f59de6678c68e8fe843a9fa601f606870347da6a587963c95e8869902be",
  "source_column_headers_explicit": false
}
Complete studied output field

additional_extractors.newspaper4k.html

<div><h1>Protocole hors ligne de conservation des relations</h1><p>Ce document synth&#233;tique sert &#224; comparer deux balisages &#224; texte, ordre, classes et environnement identiques. Les donn&#233;es sont fictives et ne d&#233;crivent ni Edikka ni un client.</p><h2>Mesures synth&#233;tiques</h2>Mesures synth&#233;tiquesIndicateurUnit&#233;P&#233;riodeValeurTemps de r&#233;ponsesecondesmai 20261,8<p>La conclusion doit &#234;tre relue avec son unit&#233;, sa p&#233;riode, son p&#233;rim&#232;tre et sa r&#233;serve. La pr&#233;sence d&#8217;un mot dans une sortie ne suffit pas &#224; prouver que la relation a &#233;t&#233; conserv&#233;e.</p></div>
Archived version and options

newspaper4k 0.9.3.1 · Execution 2026-09-20

{
  "version": "0.9.3.1",
  "configuration": "No configuration field in this output; consult the archived runner."
}

Archived runner · Python options

SHA-256 hashes
HTML input
1e0a3b0807f57d794396fce3fe07c0259b8c3c4c40be85cea27483f2833516b1
Evaluated output field
83fe1f59de6678c68e8fe843a9fa601f606870347da6a587963c95e8869902be

Complete raw output file ↗

Identical text and isolated intervention within this pair, according to the protocol witnesses. Translations are not independent observations. View pair witnesses.

LOOK FURTHER

Every result has a scope.

These observations do not rank tools. They do not measure overall accessibility, search ranking, or understanding or citation by AI.

Claims excluded by the protocol