chaffjs 0.19.0 → 0.21.0

package/rules/orthographic-variant.yaml

@@ -0,0 +1,90 @@
+id: orthographic-variant
+layer: L3
+status: stable
+how_to_find: orthographic-variant
+# 語の一覧は持たない。文書の中の語を、日本語は辞書の形の読み（出来る・できる はどちらもデキル）で、カタカナ語は小さい字と「ー」を
+# ならした形で、英字は大文字小文字とハイフンを外した形で束ね、同じ語の書き方が二つあれば少ないほうを指す。
+# 読みが同じで意味の違う語（成る・なる）は lang-ja の orthographic-variant-distinct に並べて外す。
+uses: [pos]
+severity: info
+
+name:
+  ja: 同じ語の書き方がゆれている
+  en: One word spelled two ways
+
+why:
+  ja: 「出来る」と「できる」、「引っ越し」と「引越し」、「ウィンドウ」と「ウインドウ」、「e-mail」と「email」のように、同じ語を一つの文書の中で二通りに書くと、検索で別の語になり、読み手は別のものを指しているのかと迷います。どちらが正しいかは決めません。chaff は語の一覧を持たず、文書の中の語を読み（日本語）やならした綴り（カタカナ語と英字）で束ね、文書の中で少ないほうを指摘します。補助動詞と形式名詞（見て下さい、その事）は、本来の使い方（資料を下さい）と別に比べるので、「補助動詞はかなで書く」と決めた文書はそろっていると見ます。漢字の違う同音の語（書く・描く、機会・機械）は別の語と見ます。少ないほうが三分の一を超える文書は、使い分けていると見て言いません。
+  en: Writing one word two ways in one document (e-mail and email, GitHub and Github) makes search treat them as different words and leaves the reader wondering whether two things are meant. chaff does not decide which way is right and keeps no list of words; it groups the document's own words, by reading in Japanese and by spelling without hyphens and case in English, and points at whichever way the document uses less. The case of the first letter is left out, since a sentence or a title gives it (Email, email), and so are words in capitals only. It says nothing when the minority is more than a third, since the document then uses both on purpose.
+
+message:
+  ja: 「{written}」と書いています（この文書はふつう「{other}」と書く語です。{of} 箇所のうち {count} 箇所が違う）
+  en: '"{written}" here, where the document usually writes "{other}" ({count} of {of})'
+
+how_to_fix:
+  ja: 文書の多いほうの書き方に合わせてください。チームで決めているなら、chaff.yaml の prefer に書くと、それからは preferred-term がその向きで見ます。
+  en: Match the way the rest of the document writes it. If the team has decided, list it under prefer in chaff.yaml, and preferred-term checks that direction from then on.
+
+placeholders:
+  written:
+    ja: この箇所の書き方（日本語は辞書の形）
+    en: the spelling here
+  other:
+    ja: 文書でいちばん多い書き方
+    en: the spelling the document uses most
+
+# 少ないほうが語の何パーセントまでなら「ゆれている」と見るか。これを超えたら、文書が使い分けている。
+levels: { strict: 49, normal: 34, relaxed: 20 }
+
+use_for: [business, technical, blog, legal, docs, academic, literature, speech]
+off_for:
+  literature: "A narrator may hedge, speak with ら dropped, and spell as the work's time and place did."
+
+group: consistency
+summary:
+  ja: 「出来る」と「できる」、「引っ越し」と「引越し」のように、同じ語を一つの文書の中で二通りに書いている所（語の一覧なしで見つけます）
+  en: One word written two ways in one document (e-mail and email, GitHub and Github), found without a word list
+example:
+  ja:
+    before: |
+      申請は窓口でできます。郵送でもできます。代理の人も申請出来ます。
+    after: |
+      申請は窓口でできます。郵送でもできます。代理の人も申請できます。
+  en:
+    before: |
+      Send the form by email. We answer every email within a day. Attach the receipt to the e-mail.
+    after: |
+      Send the form by email. We answer every email within a day. Attach the receipt to the email.
+
+not_flagged:
+  ja: 読みが同じでも漢字の違う語（書く・描く）、補助動詞・形式名詞と本来の使い方の書き分け（見て下さい・資料を下さい）、書き分けが意味の違いになる語（成る・なる、言う・いう）、語末の「ー」だけの違い（katakana-long-vowel が見ます）、鉤括弧や引用符の中、chaff.yaml の names に書いた名前は見ません。英字は、頭の字の大文字小文字（Email と email）と、大文字だけの語（強調や略語）を書き方の違いにしません。少ないほうが三分の一を超えるときも言いません。文学のジャンルでは既定で止めています（その時代の書き方のため）。
+  en: The case of a word's first letter (Email, email; Long-Term, long-term) is not a spelling, nor is a word in capitals only (emphasis or an acronym). Hyphenated words that are another word (re-sign, resign), part of an address (github.com), anything in quotation marks and the team's names are not compared. Nothing is reported when the minority is more than a third. The literature preset turns it off, since old spelling belongs to its time.
+level_meaning:
+  ja: 少ないほうが語の {limit}% 以下のとき指摘
+  en: reported when the minority spelling is {limit}% of the word's uses or less
+rewrite:
+  depth: light
+  ja:
+    direction: 同じ語の書き方を、文書でいちばん多い書き方にそろえてください。意味の違う語に書き分けているなら、そのままにします。
+    pairs:
+      - before: 代理の人も申請出来ます。
+        after: 代理の人も申請できます。
+      - before: 引越しの日を決めてください。
+        after: 引っ越しの日を決めてください。
+    keep:
+      - 語の意味と、文の中の位置
+      - 意味の違いで書き分けている語（資料を下さい・見てください）
+    avoid:
+      - 少ないほうに合わせて、多いほうを書き換える
+      - 引用した文の書き方を変える
+  en:
+    direction: Write each word the way the rest of the document writes it. If two spellings mark two meanings, leave them.
+    pairs:
+      - before: Attach the receipt to the e-mail.
+        after: Attach the receipt to the email.
+      - before: The code is on Github.
+        after: The code is on GitHub.
+    keep:
+      - the meaning and the place of the word
+      - quoted text as quoted
+    avoid:
+      - changing the majority to match the minority

package/rules/outside-announced-list.yaml

@@ -1,6 +1,6 @@
 id: outside-announced-list
 layer: L3
-status: experimental
+status: stable
 how_to_find: outside-announced-list
 requires: [pos]
 severity: warning
@@ -56,6 +56,7 @@ not_flagged:
   ja: 一覧を言う語（対応OSは、に対応、版）のすぐ隣に並べた固有名詞だけを読みます。最初に二つ以上の名前を挙げた文を一覧とし、それより前の文は見ません。組の違う語（対応する OS と、対応する言語）は比べません。
   en: Only proper nouns right next to a listing word (supported on, available in, version) are read. The first sentence naming two or more is the list, and earlier sentences are not checked. Words of different groups (platforms and languages) are not compared.
 rewrite:
+  depth: light
   ja:
     direction: 一覧と後の文のどちらが正しいかを書き手に確かめてください。後の文が正しければ一覧に名前を足し、一覧が正しければ後の文を直すか消します。
     pairs:

package/rules/oxford-comma-consistency.yaml

@@ -1,8 +1,10 @@
 id: oxford-comma-consistency
 layer: L3
-status: experimental
+status: stable
 how_to_find: oxford-comma
 languages: [en]
+# The list, its items and its conjunction are read on tokens; without them there is nothing to judge.
+requires: [pos]
 extra_word_lists: [participle-word, example-marker, pair-opener]
 severity: info
 

package/rules/padded-intro.yaml

@@ -30,6 +30,12 @@ levels:
 
 # プレスリリース・メール・議事録の冒頭は理由や用件を言う所で、つかみではない。
 use_for: [blog, business/report, business/proposal, legal, docs, academic, literature, speech]
+off_for:
+  legal: "Legal drafting repeats its terms and its sentence heads on purpose, puts the actor in the passive (shall be deemed), and joins long noun compounds (個人情報取扱事業者). This rule would only flag the form."
+  docs: "A help page repeats its heading in its first line (To upload a file, ...) and states no claims to back up."
+  academic: "Science writes in the passive, hedges its claims on purpose (may, suggest) and cites numbers as its evidence."
+  literature: "Length, rhythm, repetition and register are the author's choices in literature, and dialogue mixes styles. Only what one work can hold itself to runs: consistent spacing and commas, and numbering without gaps."
+  speech: "A speech opens with greetings and closes with thanks: that is its form, not padding."
 
 group: wording
 summary:
@@ -50,6 +56,7 @@ level_meaning:
   ja: 当てはまる書き出しが {limit} 箇所あれば指摘
   en: reported once there are {limit}
 rewrite:
+  depth: light
   ja:
     direction: どの文書にも当てはまる書き出し（「近年、〜が注目されています」）を消し、この文書だけが扱う状況か主張から始めてください。それが本文に無ければ書き手に聞きます。
     pairs:

package/rules/paragraph-length-variance.yaml

@@ -1,6 +1,6 @@
 id: paragraph-length-variance
 layer: L1
-status: experimental
+status: stable
 how_to_find: paragraph-variance
 severity: info
 
@@ -31,6 +31,12 @@ options:
       en: The fewest paragraphs measured. With few paragraphs, the variation is small by chance, and a person's writing looks uniform too
 
 use_for: [blog, legal, docs, academic, literature, speech]
+off_for:
+  legal: "Legal drafting repeats its terms and its sentence heads on purpose, puts the actor in the passive (shall be deemed), and joins long noun compounds (個人情報取扱事業者). This rule would only flag the form."
+  docs: "A help page repeats its heading in its first line (To upload a file, ...) and states no claims to back up."
+  academic: "Science writes in the passive, hedges its claims on purpose (may, suggest) and cites numbers as its evidence."
+  literature: "Length, rhythm, repetition and register are the author's choices in literature, and dialogue mixes styles. Only what one work can hold itself to runs: consistent spacing and commas, and numbering without gaps."
+  speech: "A speech opens with greetings and closes with thanks: that is its form, not padding."
 
 group: ai-tells
 summary:

package/rules/paragraph-restatement.yaml

@@ -0,0 +1,101 @@
+id: paragraph-restatement
+layer: L2
+status: stable
+how_to_find: paragraph-restatement
+requires: [pos]
+severity: info
+
+name:
+  ja: 前の段落を言い直すだけの段落
+  en: A paragraph that only restates the one before
+
+why:
+  ja: 「つまり」「このように」「言い換えれば」で始まる段落が、すぐ前の段落と同じ語ばかりで書かれていると、読み手は同じことを二度読むことになります。まとめは、前の段落に無いこと（結論、次に何をするか、数字）を足してはじめて役に立ちます。chaff は、言い直しを告げる語（語彙表）で始まる段落の内容語（名詞・動詞・形容詞・副詞・数）のうち、前の段落にすでにある語の割合を数えます。
+  en: When a paragraph opens with "In other words" or "In short" and then uses mostly the words of the paragraph before it, the reader reads the same thing twice. A summary helps only when it adds something the paragraph before did not say (the conclusion, what to do next, a figure). chaff counts, for a paragraph that opens with a restating word (from the lexicon), the share of its content words (nouns, verbs, adjectives, adverbs, numbers) already in the paragraph before.
+
+message:
+  ja: 「{opener}」で始まるこの段落は、内容語の {share}% が前の段落の繰り返しです
+  en: 'This paragraph opens with "{opener}" and {share}% of its content words repeat the paragraph before'
+
+how_to_fix:
+  ja: 段落を消して、前の段落で言い切ってください。まとめが要るなら、前の段落に無いこと（結論、次の一手、数字）を書きます。
+  en: Cut the paragraph and let the one before stand. If a summary is needed, make it say what the paragraph before did not (the conclusion, the next step, a figure).
+
+# 言い直しを告げる語で始まる段落の内容語のうち、前の段落にある語の割合（パーセント）。これ以上で指摘。
+levels: { strict: 40, normal: 50, relaxed: 60 }
+
+use_for: [blog, business, technical, docs, academic, legal, literature, speech]
+off_for:
+  legal: "すなわち and that is, open a definition that restates on purpose."
+  literature: "Length, rhythm, repetition and register are the author's choices in literature, and dialogue mixes styles. Only what one work can hold itself to runs: consistent spacing and commas, and numbering without gaps."
+  speech: "A speaker says it again so the room hears it: 要するに, in other words."
+
+group: readability
+summary:
+  ja: 「つまり」「このように」で始まり、前の段落の語をほとんど繰り返すだけの段落
+  en: A paragraph that opens with "In other words" or "In short" and mostly repeats the paragraph before
+
+example:
+  ja:
+    config: { genre: blog/tech }
+    before: |
+      新しい在庫システムでは、店舗の在庫を一つの画面で数えられます。棚卸しの時間も短くなります。
+
+      このように、新しい在庫システムでは店舗の在庫を一つの画面で数えられ、棚卸しの時間が短くなります。
+    after: |
+      新しい在庫システムでは、店舗の在庫を一つの画面で数えられます。棚卸しの時間も短くなります。
+
+      来月から全店で使い始め、最初の棚卸しは 4 月 10 日です。
+  en:
+    config: { genre: blog/tech }
+    before: |
+      The new inventory system counts every store's stock on one screen, and a stocktake takes less time.
+
+      In other words, the new inventory system counts the stock of every store on one screen, so a stocktake takes less time.
+    after: |
+      The new inventory system counts every store's stock on one screen, and a stocktake takes less time.
+
+      Every store switches next month, and the first stocktake is on April 10.
+
+not_flagged:
+  ja: 言い直しを告げる語で始まらない段落、内容語が 4 つに満たない段落、見出し・箇条書き・表・コードを挟んだ段落は見ません。前の段落の語の割合だけを見るので、まとめが前の数段落を受けていても、すぐ前の段落と比べます。法務・文学・話し言葉のジャンルでは見ません。
+  en: 'A paragraph that does not open with a restating word, one with fewer than four content words, and one separated from the paragraph before by a heading, a list, a table or code are not checked. Words that open other sentences too ("That is why", "Again the drums") count only with a comma after them ("That is,"). Legal, literary and speech genres do not run this rule.'
+level_meaning:
+  ja: 内容語の {limit}% 以上が前の段落の繰り返しのとき指摘
+  en: reported when {limit}% or more of the paragraph's content words repeat the paragraph before
+rewrite:
+  depth: structure
+  ja:
+    direction: 言い直しの段落を消すか、前の段落に無いことを書く段落に変えてください。結論・次にすること・数字のうち、本文にあるものを書きます。「つまり」「このように」で前の段落をなぞるだけの文は残しません。
+    pairs:
+      - before: |
+          このように、新しい在庫システムでは店舗の在庫を一つの画面で数えられ、棚卸しの時間が短くなります。
+        after: |
+          来月から全店で使い始め、最初の棚卸しは 4 月 10 日です。
+      - before: |
+          つまり、申込書は月末までに出す必要があり、遅れると受け付けられません。
+        after: |
+          郵送の場合は、月末の消印まで有効です。
+    keep:
+      - 前の段落の中身
+      - 本文にある日付・数字・名前
+    avoid:
+      - 本文に無い事実を足す
+      - 前の段落を別の言い方でもう一度書く
+  en:
+    direction: Cut the restating paragraph, or turn it into one that says what the paragraph before did not. Use the conclusion, the next step or a figure the text gives. Do not keep a sentence that only walks over the previous paragraph again after "In other words".
+    pairs:
+      - before: |
+          In other words, the new inventory system counts the stock of every store on one screen, so a stocktake takes less time.
+        after: |
+          Every store switches next month, and the first stocktake is on April 10.
+      - before: |
+          In short, the form must reach us by the end of the month, or it will not be accepted.
+        after: |
+          A posted form counts if it is postmarked by the last day of the month.
+    keep:
+      - what the paragraph before says
+      - the dates, figures and names in the text
+    avoid:
+      - adding a fact the text does not have
+      - saying the paragraph before again in other words

package/rules/percent-sum-mismatch.yaml

@@ -1,6 +1,6 @@
 id: percent-sum-mismatch
 layer: L3
-status: experimental
+status: stable
 how_to_find: percent-sum-mismatch
 requires: [quantities]
 severity: warning
@@ -24,6 +24,8 @@ how_to_fix:
 levels: { strict: error, normal: warning, relaxed: info }
 
 use_for: [business, technical, blog, legal, docs, academic, literature, speech]
+off_for:
+  literature: "Length, rhythm, repetition and register are the author's choices in literature, and dialogue mixes styles. Only what one work can hold itself to runs: consistent spacing and commas, and numbering without gaps."
 
 group: facts
 summary:

package/rules/person-consistency.yaml

@@ -0,0 +1,62 @@
+id: person-consistency
+layer: L2
+status: stable
+how_to_find: person-mix
+word_list: person-word
+extra_word_lists: [person-form]
+severity: info
+
+name:
+  ja: 自分や読み手の呼び方が揃っていない
+  en: The document names itself or its reader two ways
+
+why:
+  ja: 文書が自分を「私たち」と呼ぶか「弊社」「当社」と呼ぶか、書き手を「私」と書くか「筆者」と書くか、英語で読み手を you と呼ぶか one と呼ぶかは文体の選択で、どれが正しいかは決めません。一つの文書の中で混ざると、読み手は話し手が変わったのか、別の誰かのことなのかを考えます。継ぎ合わせた文書や、別の人が足した段落によく起きます。chaff は呼び方ごとに文書の中で多いほうを文書の書き方とし、少ないほうを指します。鉤括弧と引用符の中、引用の段落、コードは数えません。英語の we と I は、多くの文書で別の相手（組織と話し手）を指すので比べません。
+  en: Whether a Japanese document calls itself 私たち or 弊社 and its writer 私 or 筆者, and whether an English document addresses its reader as "you" or as "one", is a choice of voice, and chaff does not pick one. Mixed in one document, the reader wonders whether the speaker changed or someone else is meant; it happens most where text was stitched together or another writer added a paragraph. chaff takes the form the document uses most as its habit and points at the other. Quotations, quoted blocks and code are not counted. English "we" and "I" are not compared, since in most documents they name two parties (the organisation and the speaker).
+
+message:
+  ja: 「{matched}」と書いています（この文書はふつう「{usual}」。{of} 箇所のうち {count} 箇所が違う）
+  en: '"{matched}" here, where the document usually says "{usual}" ({count} of {of})'
+
+placeholders:
+  matched:
+    ja: 見つけた呼び方
+    en: the form found
+  usual:
+    ja: 文書でふつうの呼び方
+    en: the form the document usually uses
+
+how_to_fix:
+  ja: 同じ相手を指しているなら、文書の多いほうの呼び方に揃えてください。別の相手（会社と、書き手個人）を指しているなら、その違いが読み手に分かるように書きます。
+  en: If the same party is meant, use the form the rest of the document uses. If another party is meant (the company and the writer), make the difference clear to the reader.
+
+# 少ないほうの呼び方が、その呼び方全体の何パーセントまでなら混ざりと見るか。超えたら、文書が使い分けている。
+levels: { strict: 30, normal: 20, relaxed: 10 }
+
+use_for: [technical, blog, business, legal, docs, academic, literature, speech]
+off_for:
+  literature: "A story's narrator and its characters each say 私 or 僕 in their own voice."
+  speech/transcript: "Several people speak, each naming themselves their own way."
+
+group: consistency
+summary:
+  ja: 「私たち」と「弊社」、「私」と「筆者」のように、一つの文書が自分や読み手を二通りに呼んでいる所
+  en: 'One document naming itself two ways (私たち and 弊社) or its reader two ways ("you" and "one")'
+example:
+  ja:
+    before: |
+      弊社は、来月から新しい窓口を開きます。弊社の担当者が、平日の午前に受け付けます。私たちは、相談の内容を記録に残します。弊社は、記録を一年間保管します。保管した記録は、弊社の外には出しません。
+    after: |
+      弊社は、来月から新しい窓口を開きます。弊社の担当者が、平日の午前に受け付けます。弊社は、相談の内容を記録に残します。弊社は、記録を一年間保管します。保管した記録は、弊社の外には出しません。
+  en:
+    before: |
+      You can book a slot online. Your booking shows on the next page. If you miss it, you can book again. One can also call the desk.
+    after: |
+      You can book a slot online. Your booking shows on the next page. If you miss it, you can book again. You can also call the desk.
+
+not_flagged:
+  ja: 呼び方が一通りの文書と、少ないほうが決まった割合を超える文書（使い分けていると見ます）は指摘しません。多いほうが 3 箇所に満たない文書も見ません。鉤括弧と引用符の中、引用の段落、コード、前後に漢字やカタカナが続く語（東京本社、当社比）は数えません。英語の one は、can や must の前の主語の one だけを数え、no one や the one や前置詞の後ろの one は数えません。どの呼び方が正しいかは決めません。
+  en: A document with one form, or whose other form passes a set share of the places (it uses both on purpose), is not reported, nor one whose usual form appears fewer than 3 times. Quotations, quoted blocks, code and words joined to longer ones (東京本社) are not counted. Only "one" as the subject of a modal verb is counted, not "no one", "the one" or "one" after a preposition. Which form is right is not decided.
+level_meaning:
+  ja: 少ないほうの呼び方が全体の {limit} パーセント以下のとき指摘
+  en: reported when the other form is {limit} percent of the places or fewer

package/rules/preamble-length.yaml

@@ -24,6 +24,12 @@ how_to_fix:
 levels: { strict: 1, normal: 2, relaxed: 4 }
 
 use_for: [business, legal, docs, academic, literature, speech]
+off_for:
+  legal: "Legal drafting repeats its terms and its sentence heads on purpose, puts the actor in the passive (shall be deemed), and joins long noun compounds (個人情報取扱事業者). This rule would only flag the form."
+  docs: "A help page repeats its heading in its first line (To upload a file, ...) and states no claims to back up."
+  academic: "Science writes in the passive, hedges its claims on purpose (may, suggest) and cites numbers as its evidence."
+  literature: "Length, rhythm, repetition and register are the author's choices in literature, and dialogue mixes styles. Only what one work can hold itself to runs: consistent spacing and commas, and numbering without gaps."
+  speech: "A speech opens with greetings and closes with thanks: that is its form, not padding."
 
 group: structure
 summary:

package/rules/preferred-term.yaml

@@ -1,6 +1,6 @@
 id: preferred-term
 layer: L2
-status: experimental
+status: stable
 how_to_find: preferred-term
 word_list: preferred-term
 severity: warning

package/rules/proper-noun-density.yaml

@@ -1,6 +1,6 @@
 id: proper-noun-density
 layer: L3
-status: experimental
+status: stable
 how_to_find: proper-noun-density
 requires: [pos]
 # chaff.yaml の names に並べた名前は、何語に割れても 1 つと数える。名前そのものは数える（読み手が知らないことは変わらない）。
@@ -29,6 +29,13 @@ levels:
   en: { strict: 150, normal: 200, relaxed: 260 }
 
 use_for: [blog, business, technical, legal, docs, academic, literature, speech]
+off_for:
+  legal: "Legal drafting repeats its terms and its sentence heads on purpose, puts the actor in the passive (shall be deemed), and joins long noun compounds (個人情報取扱事業者). This rule would only flag the form."
+  docs: "A help page repeats its heading in its first line (To upload a file, ...) and states no claims to back up."
+  academic: "Science writes in the passive, hedges its claims on purpose (may, suggest) and cites numbers as its evidence."
+  literature: "Length, rhythm, repetition and register are the author's choices in literature, and dialogue mixes styles. Only what one work can hold itself to runs: consistent spacing and commas, and numbering without gaps."
+  speech: "A speech opens with greetings and closes with thanks: that is its form, not padding."
+  business/press-release: "A release names the agency, the people quoted, the mission and the place: that is its news, not a crowd of names to thin out. Every English release in the corpus (GSA, NASA, NOAA) is over the limit."
 
 group: readability
 summary:

package/rules/quote-style-consistency.yaml

@@ -0,0 +1,66 @@
+id: quote-style-consistency
+layer: L1
+status: stable
+how_to_find: mark-consistency
+word_list: quote-style
+languages: [en]
+severity: info
+
+name:
+  ja: 引用符の直と曲の混在
+  en: Straight and curly quotes mixed
+
+why:
+  ja: 英語の引用符とアポストロフィを、直の形（" '）と曲がった形（“ ” ’）で混ぜて書くと、別々の所から貼り合わせた文書に見え、検索や置換でも片方を取りこぼします。どちらが正しいかは決めません。文書の中で少ないほうを指します。
+  en: Straight quotes and apostrophes (" ') mixed with curly ones (“ ” ’) make a document look pasted together, and a search or replace finds only one kind. chaff does not pick a side; it points at the less common way in the document.
+
+message:
+  ja: 引用符を {mark} と書いています（この文書はふつう {usual}。{of} 箇所のうち {count} 箇所が違う）
+  en: The quote mark {mark} here, where the document usually writes {usual} ({count} of {of})
+
+# 少ないほうが 5 箇所以上で、全体の 2 割を超えるとき。書き損じではなく書き方が二通りあるので、1 件にまとめて両方の数を言う。
+messages:
+  mixed:
+    ja: "引用符を、曲がった形で {spaced} 箇所、直の形で {touching} 箇所書いています。この文書には書き方が二通りあります"
+    en: "Quote marks are written two ways in this document ({spaced} curly, {touching} straight)"
+
+how_to_fix:
+  ja: 文書の多いほうの形に合わせてください。
+  en: Match the form the rest of the document uses.
+
+levels: { strict: 1, normal: 1, relaxed: 3 }
+
+use_for: [business, technical, blog, legal, docs, academic, literature, speech]
+off_for:
+  docs: measured by yarn rules:measure
+
+group: consistency
+summary:
+  ja: 英語で、直の引用符（" '）と曲がった引用符（“ ” ’）が一つの文書の中で混ざっている所
+  en: Straight quotes (" ') and curly quotes (“ ” ’) mixed in one document
+example:
+  en:
+    before: |
+      The "draft" is ready. The "final" is due. The “review” is next.
+    after: |
+      The "draft" is ready. The "final" is due. The "review" is next.
+
+not_flagged:
+  ja: 英字に接する引用符だけを数えます。コードの中と、二重引用符と一重引用符（アポストロフィ）の間の違いは見ません（種類ごとに数えます）。
+  en: Only marks touching a letter are counted, and code is skipped. Double and single quotes (apostrophes) are counted as separate kinds.
+level_meaning:
+  ja: 少ないほうの書き方が {limit} 箇所あれば指摘
+  en: reported once the less common way appears {limit} times
+rewrite:
+  depth: light
+  en:
+    direction: Change the marked quotes to the form the rest of the document uses. Change nothing else.
+    pairs:
+      - before: The “review” is next.
+        after: The "review" is next.
+      - before: The team’s plan holds.
+        after: The team's plan holds.
+    keep:
+      - the words inside the quotes
+    avoid:
+      - changing the quotes the document already writes the common way

package/rules/ra-nuki.yaml

@@ -1,6 +1,6 @@
 id: ra-nuki
 layer: L2
-status: experimental
+status: stable
 how_to_find: ra-nuki
 # 一段動詞かどうかは品詞解析で読む。品詞が無ければ動かない。
 requires: [pos]
@@ -26,6 +26,9 @@ how_to_fix:
 levels: { strict: error, normal: warning, relaxed: info }
 
 use_for: [business, technical, blog, legal, docs, academic, literature, speech]
+off_for:
+  literature: "A narrator may hedge, speak with ら dropped, and spell as the work's time and place did."
+  speech: "A transcript keeps the speaker's words: hedges and ら dropped as spoken."
 
 group: wording
 summary:

package/rules/redundant-expression.yaml

@@ -0,0 +1,79 @@
+id: redundant-expression
+layer: L2
+status: stable
+how_to_find: phrase-match
+word_list: redundant-expression
+uses: [pos]
+severity: info
+
+name:
+  ja: 重言
+  en: Redundant pair
+
+why:
+  ja: 「頭痛が痛い」「一番最初」「後で後悔する」のように同じ意味の語を重ねると、読み手は二度目の語に意味を探してつまずき、書き手が言葉を選んでいないように読めます。重ねを外しても意味は変わりません。言い回しの一覧は言語ごとの語彙表にあり、それぞれに重ねを外した形が付いています。辞書が一語として載せる慣用（排気ガス、射程距離）は数えません。
+  en: Two words that say the same thing ("end result", "each and every", "revert back") make the reader look for a meaning in the second word that is not there, and read as unedited. Dropping the repeat changes nothing. The pairs are listed in each language's lexicon, each with the form to use instead. Idioms that are sometimes needed ("whether or not", "period of time") are not counted.
+
+message:
+  ja: 「{matched}」は同じ意味の語を重ねています
+  en: '"{matched}" says the same thing twice'
+
+how_to_fix:
+  ja: 重ねた語の一方を消してください（一番最初 → 最初、後で後悔する → 後悔する）。
+  en: Drop the repeat ("end result" → "result", "revert back" → "revert").
+
+# 何回あれば指摘するか。1 は一つ目から。
+levels: { strict: 1, normal: 1, relaxed: 2 }
+
+use_for: [business, technical, blog, legal, docs, academic, literature, speech]
+off_for:
+  literature: "Length, rhythm, repetition and register are the author's choices in literature, and dialogue mixes styles. Only what one work can hold itself to runs: consistent spacing and commas, and numbering without gaps."
+
+group: wording
+summary:
+  ja: 「頭痛が痛い」「一番最初」「後で後悔」のような、同じ意味の語の重なり
+  en: Two words that say the same thing, such as "end result" or "each and every"
+example:
+  ja:
+    before: |
+      一番最初に、手順を確認してください。
+    after: |
+      最初に、手順を確認してください。
+  en:
+    before: |
+      The end result is a shorter build.
+    after: |
+      The result is a shorter build.
+
+not_flagged:
+  ja: 辞書が一語として載せる慣用（排気ガス、射程距離、製造メーカー）、重ねて強める書き方として定着したもの（犯罪を犯す、あらかじめ予約）は指しません。文学のジャンルでは既定で止めています。
+  en: Idioms that are sometimes needed ("whether or not", "period of time", "past experience", "plan ahead") are not reported. The literature preset turns it off.
+level_meaning:
+  ja: 重言を含む文が {limit} 文あれば指摘
+  en: reported once {limit} sentences hold one
+rewrite:
+  depth: light
+  ja:
+    direction: 重ねた語の一方を消して、一度だけ言ってください。語彙表の語ごとの書き換え（一番最初 → 最初）に従い、文のほかの所は変えません。
+    pairs:
+      - before: 一番最初に、手順を確認してください。
+        after: 最初に、手順を確認してください。
+      - before: 返事を返すまでに三日かかりました。
+        after: 返事をするまでに三日かかりました。
+    keep:
+      - 文の意味と語順
+      - 重ねていない語
+    avoid:
+      - 重ねを外すついでに文を書き換える
+  en:
+    direction: Drop one of the two words that say the same thing, following the lexicon's replacement for the phrase ("end result" → "result"). Leave the rest of the sentence alone.
+    pairs:
+      - before: The end result is a shorter build.
+        after: The result is a shorter build.
+      - before: We reverted back to the old schema.
+        after: We reverted to the old schema.
+    keep:
+      - what the sentence says
+      - the words around the pair
+    avoid:
+      - rewriting the whole sentence while dropping the repeat