Found while doing a proofreading pass over the whole book (#typo-pass-2026-09-02). These are content-correctness questions in the Korpusanalyse (readability-index) chapter, not simple typos — each requires deciding what the "correct" number or explanation should be, so I'm listing them here instead of silently rewriting worked examples. Grouped by sub-topic.
1. Worked-example arithmetic doesn't match the stated result
Both in corpus_analysis/corpus-analysis_text_complexity.md.
Flesch example (section "Flesch-Lesbarkeitsindex", "Beispiel"): the syllable breakdown sums to 24 syllables / 13 words → ASW = 1.846..., and the formula line itself shows 58,5 × 1.84 — but the stated final result is 61.7. Recomputing 180 - 13 - (58.5 × 1.84) gives 59.4, not 61.7. This isn't just cosmetic: 61.7 falls in the "60-70: Durchschnittlich" band while 59.4 falls in "50-60: Anspruchsvoll" — so the example's own classification sentence ("Der Satz wird als durchschnittlich eingestuft") would also need to change if the number is corrected.
Coleman-Liau example (section "Coleman-Liau Score", "Beispiel"): with the stated inputs L=6,31 and S=0.077, 0,0588 × 6,31 × 100 - 0,296 × 0,077 × 100 - 15,8 computes to 19.02, not the stated 18,52. This doesn't change the stated classification ("sehr schwierig" still holds either way).
Wiener example (same file): recomputed correctly and matches the stated 3.32 — no arithmetic error here, but FYI the result falls below the chapter's own interpretation scale, which starts at "4-5: Sehr einfach" — worth double-checking whether the example sentence is meant to score that low, or whether an input was mistyped.
I'd guess someone hand-computed these and made an arithmetic slip, or an earlier version of the example sentence had different word/syllable counts and the walkthrough wasn't fully updated — but I don't want to guess which number is "true" and silently rewrite a worked example's conclusion.
2. ARI / Coleman-Liau interpretation ceiling: "13" vs "14"
corpus-analysis_text_complexity.md's own ARI table caps at 13+: 18+ Jahre (Hochschule). But several other places in the book consistently say the ceiling is 14 instead:
corpus-analysis_assessment.md (Frage 4 feedback): "ARI (1-14+)"corpus-analysis_analysis.ipynb: the ARI cell says "1 (sehr leicht) bis 14 (schwierig)", a later cell says "Score über 14 außerhalb der festgelegten Skala", and the Coleman-Liau cell (which per the .md chapter should use the same "Analog zum ARI" scale) also says "1 (sehr leicht) bis 14 (schwierig)"
That same assessment.md feedback also says "Flesch (0-180)", which doesn't match the chapter's own Flesch table (0-100) at all — my guess is this confuses the formula's constant (180 - ASL - ...) with the score's practical range, but I left it alone since it's adjacent to the same cluster.
Given how consistently "14" (and "0-180") appear across the notebook and assessment quiz, this doesn't look like a one-off typo — it looks like the notebook and the .md chapter may have been written from different assumptions about the scale. Worth deciding which is authoritative and aligning the rest.
3. Wiener Sachtextformel described as dictionary-based, but its own formula isn't
corpus-analysis_assessment.md (Frage 1, option 4 feedback): "Die Wiener Sachtextformel ist wörterbuch-basiert, andere Indexe arbeiten über Länge beziehungsweise Silben."corpus-analysis_resumee.md: "Wiener Sachtextformel (auf Deutsch ausgelegt, operiert auf Satzlänge und Anteil von schwierigen Wörtern)"
But the chapter's own formula definition (text_complexity.md) computes Wiener from MS (% of words with 3+ syllables), ASL (avg. sentence length), and IW (count of words >6 letters) — no dictionary lookup at all. None of the four formulas covered in this chapter are dictionary-based (that's the "schwierige Wörter" method described separately, earlier in the chapter, as a different, non-index-based approach). These two spots seem to describe Wiener inaccurately; fixing them means rewriting an explanation, not swapping a word, so I left them for a content review pass.
4. Assessment quiz: self-contradictory feedback on the "3-syllable" edge case
corpus-analysis_assessment.md, Frage 1: the chapter itself is ambiguous about where the syllable cutoff falls (text_complexity.md line 28 says "Wörter mit ein oder zwei Silben gelten als leicht, Wörter mit mehr als drei Silben gelten als schwierig" — leaving exactly-3-syllable words undefined). This ambiguity surfaces as a contradiction in the quiz: option 2's own answer text says "mehr als drei Silben gelten... als schwer" (>3 = hard) while its feedback says "3 und mehr Silben gelten als schwer" (≥3 = hard). Worth deciding the actual cutoff and fixing both the chapter definition and the quiz consistently.
5. Musterlösung arithmetic: "5 Punkte" vs. the scenario's own numbers
corpus-analysis_assessment.md, Frage 7 Musterlösung: states "Durchschnittlich ca. 5 Punkte pro 5-Jahres-Periode", but the scenario's own stated score ranges (70-75 → 60-65 → 50-55) imply a decline of roughly 10 points per 5-year period (midpoints 72.5 → 62.5 → 52.5), not 5.
6. corpus-analysis_visualization.ipynb: instructional block describes a workflow that isn't in this notebook
The second "Informationen zum Ausführen des Notebooks" markdown cell (near the top, above the "Ergebnisdatei laden" section) describes downloading text files into a data/txt/ folder and adjusting a text_path variable — but this notebook never downloads text files or defines any such variable; it only reads a pre-computed metadata_with_readability_scores.csv into result_df. This whole paragraph looks like it was copied from a different notebook's template (the corresponding block in html_intro/html-intro_structural-analysis.ipynb does describe a real download-HTML-files step in that notebook) and never adapted. I fixed the block's plain typos and a stale GitHub link as part of the typo pass, but left the substantive "does this paragraph belong here at all" question for a content review — it may need to be rewritten to describe the actual CSV-loading step, or removed entirely.
7. Terminology drift: "Barrierearmut" vs "Barrierefreiheit"
The book's title and research question consistently define Barrierearmut as the specific operationalized concept being measured. But a few places drift to the related-but-broader term Barrierefreiheit when discussing results/implications:
corpus-analysis_resumee.md: "die Barrierefreiheit somit verringert"corpus-analysis_assessment.md Musterlösung: "Bedeutung für die Barrierefreiheit", "Negative Entwicklung für Barrierefreiheit"README.md body text also says "kommunikative Barrierefreiheit" while the title says "Barrierearmut"
Might be deliberate (Barrierearmut as the specific measure of Barrierefreiheit the broader goal), might just be loose wording. Worth a decision either way since it affects how precisely the book uses its own defined term.
8. Date range: "2011-2024" vs "2011 bis März 2025"
The book's title (front_page/intro.md and README.md, both: "(2011-2024)"), the research question (research-question_research-question.md: "im Zeitraum von 2011 bis 2024"), and the Dublin Core metadata example (corpus-collection_metadata.md: "von 2011-2024") all consistently frame the studied period as ending in 2024. But README.md's own body text says "im Zeitraum von 2011 bis März 2025" — a different end date within the very same file as the "(2011-2024)" title. (Separately, corpus-collection_building-our-corpus.md states data collection ran through mid-2025 — that's plausibly fine on its own, since scraping a year past the studied period just ensures complete 2024 coverage — but the README body text's "2025" specifically contradicts its own title.)
9. "sieben Kapitel" doesn't match the enumeration right after it
README.md: "Die OER ist in sieben Kapitel aufgeteilt: Es wird zuerst... [1] ...dann werden... [2]... und HTML... [3] erläutert, danach folgt... [4]..., die praktische Korpuserstellung wird durchgeführt [5] und abschließend wird eine quantitative Analyse... ausgeführt [6]." — the sentence claims seven but the list right after it only enumerates six items (it doesn't mention the closing Reflexion/Ausblick chapter, which would make seven if included, or the count should just say "sechs").
Found while doing a proofreading pass over the whole book (#typo-pass-2026-09-02). These are content-correctness questions in the Korpusanalyse (readability-index) chapter, not simple typos — each requires deciding what the "correct" number or explanation should be, so I'm listing them here instead of silently rewriting worked examples. Grouped by sub-topic.
1. Worked-example arithmetic doesn't match the stated result
Both in
corpus_analysis/corpus-analysis_text_complexity.md.Flesch example (section "Flesch-Lesbarkeitsindex", "Beispiel"): the syllable breakdown sums to 24 syllables / 13 words → ASW = 1.846..., and the formula line itself shows
58,5 × 1.84— but the stated final result is61.7. Recomputing180 - 13 - (58.5 × 1.84)gives 59.4, not 61.7. This isn't just cosmetic: 61.7 falls in the "60-70: Durchschnittlich" band while 59.4 falls in "50-60: Anspruchsvoll" — so the example's own classification sentence ("Der Satz wird als durchschnittlich eingestuft") would also need to change if the number is corrected.Coleman-Liau example (section "Coleman-Liau Score", "Beispiel"): with the stated inputs L=6,31 and S=0.077,
0,0588 × 6,31 × 100 - 0,296 × 0,077 × 100 - 15,8computes to 19.02, not the stated18,52. This doesn't change the stated classification ("sehr schwierig" still holds either way).Wiener example (same file): recomputed correctly and matches the stated
3.32— no arithmetic error here, but FYI the result falls below the chapter's own interpretation scale, which starts at "4-5: Sehr einfach" — worth double-checking whether the example sentence is meant to score that low, or whether an input was mistyped.I'd guess someone hand-computed these and made an arithmetic slip, or an earlier version of the example sentence had different word/syllable counts and the walkthrough wasn't fully updated — but I don't want to guess which number is "true" and silently rewrite a worked example's conclusion.
2. ARI / Coleman-Liau interpretation ceiling: "13" vs "14"
corpus-analysis_text_complexity.md's own ARI table caps at13+: 18+ Jahre (Hochschule). But several other places in the book consistently say the ceiling is 14 instead:corpus-analysis_assessment.md(Frage 4 feedback):"ARI (1-14+)"corpus-analysis_analysis.ipynb: the ARI cell says"1 (sehr leicht) bis 14 (schwierig)", a later cell says"Score über 14 außerhalb der festgelegten Skala", and the Coleman-Liau cell (which per the .md chapter should use the same "Analog zum ARI" scale) also says"1 (sehr leicht) bis 14 (schwierig)"That same assessment.md feedback also says
"Flesch (0-180)", which doesn't match the chapter's own Flesch table (0-100) at all — my guess is this confuses the formula's constant (180 - ASL - ...) with the score's practical range, but I left it alone since it's adjacent to the same cluster.Given how consistently "14" (and "0-180") appear across the notebook and assessment quiz, this doesn't look like a one-off typo — it looks like the notebook and the .md chapter may have been written from different assumptions about the scale. Worth deciding which is authoritative and aligning the rest.
3. Wiener Sachtextformel described as dictionary-based, but its own formula isn't
corpus-analysis_assessment.md(Frage 1, option 4 feedback): "Die Wiener Sachtextformel ist wörterbuch-basiert, andere Indexe arbeiten über Länge beziehungsweise Silben."corpus-analysis_resumee.md: "Wiener Sachtextformel (auf Deutsch ausgelegt, operiert auf Satzlänge und Anteil von schwierigen Wörtern)"But the chapter's own formula definition (
text_complexity.md) computes Wiener from MS (% of words with 3+ syllables), ASL (avg. sentence length), and IW (count of words >6 letters) — no dictionary lookup at all. None of the four formulas covered in this chapter are dictionary-based (that's the "schwierige Wörter" method described separately, earlier in the chapter, as a different, non-index-based approach). These two spots seem to describe Wiener inaccurately; fixing them means rewriting an explanation, not swapping a word, so I left them for a content review pass.4. Assessment quiz: self-contradictory feedback on the "3-syllable" edge case
corpus-analysis_assessment.md, Frage 1: the chapter itself is ambiguous about where the syllable cutoff falls (text_complexity.mdline 28 says "Wörter mit ein oder zwei Silben gelten als leicht, Wörter mit mehr als drei Silben gelten als schwierig" — leaving exactly-3-syllable words undefined). This ambiguity surfaces as a contradiction in the quiz: option 2's own answer text says "mehr als drei Silben gelten... als schwer" (>3 = hard) while its feedback says "3 und mehr Silben gelten als schwer" (≥3 = hard). Worth deciding the actual cutoff and fixing both the chapter definition and the quiz consistently.5. Musterlösung arithmetic: "5 Punkte" vs. the scenario's own numbers
corpus-analysis_assessment.md, Frage 7 Musterlösung: states "Durchschnittlich ca. 5 Punkte pro 5-Jahres-Periode", but the scenario's own stated score ranges (70-75 → 60-65 → 50-55) imply a decline of roughly 10 points per 5-year period (midpoints 72.5 → 62.5 → 52.5), not 5.6.
corpus-analysis_visualization.ipynb: instructional block describes a workflow that isn't in this notebookThe second "Informationen zum Ausführen des Notebooks" markdown cell (near the top, above the "Ergebnisdatei laden" section) describes downloading text files into a
data/txt/folder and adjusting atext_pathvariable — but this notebook never downloads text files or defines any such variable; it only reads a pre-computedmetadata_with_readability_scores.csvintoresult_df. This whole paragraph looks like it was copied from a different notebook's template (the corresponding block inhtml_intro/html-intro_structural-analysis.ipynbdoes describe a real download-HTML-files step in that notebook) and never adapted. I fixed the block's plain typos and a stale GitHub link as part of the typo pass, but left the substantive "does this paragraph belong here at all" question for a content review — it may need to be rewritten to describe the actual CSV-loading step, or removed entirely.7. Terminology drift: "Barrierearmut" vs "Barrierefreiheit"
The book's title and research question consistently define Barrierearmut as the specific operationalized concept being measured. But a few places drift to the related-but-broader term Barrierefreiheit when discussing results/implications:
corpus-analysis_resumee.md: "die Barrierefreiheit somit verringert"corpus-analysis_assessment.mdMusterlösung: "Bedeutung für die Barrierefreiheit", "Negative Entwicklung für Barrierefreiheit"README.mdbody text also says "kommunikative Barrierefreiheit" while the title says "Barrierearmut"Might be deliberate (Barrierearmut as the specific measure of Barrierefreiheit the broader goal), might just be loose wording. Worth a decision either way since it affects how precisely the book uses its own defined term.
8. Date range: "2011-2024" vs "2011 bis März 2025"
The book's title (
front_page/intro.mdandREADME.md, both: "(2011-2024)"), the research question (research-question_research-question.md: "im Zeitraum von 2011 bis 2024"), and the Dublin Core metadata example (corpus-collection_metadata.md: "von 2011-2024") all consistently frame the studied period as ending in 2024. ButREADME.md's own body text says "im Zeitraum von 2011 bis März 2025" — a different end date within the very same file as the "(2011-2024)" title. (Separately,corpus-collection_building-our-corpus.mdstates data collection ran through mid-2025 — that's plausibly fine on its own, since scraping a year past the studied period just ensures complete 2024 coverage — but the README body text's "2025" specifically contradicts its own title.)9. "sieben Kapitel" doesn't match the enumeration right after it
README.md: "Die OER ist in sieben Kapitel aufgeteilt: Es wird zuerst... [1] ...dann werden... [2]... und HTML... [3] erläutert, danach folgt... [4]..., die praktische Korpuserstellung wird durchgeführt [5] und abschließend wird eine quantitative Analyse... ausgeführt [6]." — the sentence claims seven but the list right after it only enumerates six items (it doesn't mention the closing Reflexion/Ausblick chapter, which would make seven if included, or the count should just say "sechs").