Re: JLRreq TF meeting notes 2023-11-24

(Resent with English translation. Japanese version follows)

JLRreq TF 2023-11-24 F2F meeting notes

attendee

atsushi, bin, kida, kobatake, murata, nat, shinyu, suzuki, tajima, taro, tlk, yamaguchi

Administrative Issues

Currently, drafts managed on Wiki are being moved to a repository, and changes will be processed via GitHub issues & PRs. The repository is under drafts. A practice file has been created at jlreq-d/sandbox/落書き帳.md, so let's practice there.

todo: Kida: Move texts from the wiki to the repository.
todo: Shimono: Create a tutorial page on the Wiki about editing methods.
Afterwards, there was a suggestion in todo: Are there better ways to manage comments and changes to draft level documents? #384 <https://github.com/w3c/jlreq/issues/384> that GitHub discussions might be a more casual approach than GitHub issues. It's currently being tried out. We'll discuss it in the next meeting.
A method is needed to indicate situations where consensus has been reached on the content, and no further discussion is necessary, but it has not yet been reflected in the documents.

Space Between Japanese and Western Characters

We are attempting to redefine this in CSS. Mr. Ishii and I are preparing a proposal to define it as a property in Unicode (to be submitted on 12/12 [UTC]).

todo: Check the content of the proposal from Mr. Ishii, discuss any differences (below in issues), and provide feedback.
todo: Separate the [spacing property] for the space around punctuation and between Kanji and Western characters, and update it reflecting the results of the discussion.
issues
Space between Japanese and Western characters: Symbols for footnotes (*†‡◊) [Combined and updated 3.3.8 f,g,h into one (3.3.8 f) #44 <https://github.com/w3c/jlreq/pull/44>]
Space between Japanese and Western characters: Emojis [span class=index in japanese text is used? #43 <https://github.com/w3c/jlreq/issues/43>]
Space between Japanese and Western characters: Treatment of combined characters (㌀㎏) [fixed typo in section 3.3.8 note #42 <https://github.com/w3c/jlreq/pull/42>]
Space between Japanese and Western characters: Treatment of circled characters [fixed English version figure 119 (img2_3_14) #41 <https://github.com/w3c/jlreq/pull/41>]
[css-text] Extra spacing between ideographs and non-fullwidth punctuation/symbols [#9479]
In Japanese books, there's no space after %
[css-text][text-autospace] Is halfwidth Kana "non-ideographic letters"? [#9471] (JLreq [[css-text][text-autospace] Is halfwidth Kana "non-ideographic letters"? #378 <https://github.com/w3c/jlreq/issues/378>])
[css-text] The definition of ideographs includes punctuation marks [#9501]
[css-text] Add black square & white square characters to the definition of ideographs? [#9603]
Discussion Points

Even for a single code point, there are cases where one might want to either create or not create space. Deciding on one can leave issues in the other usage. There should be a method to adjust the space arbitrarily. It seems easier to create space rather than remove it. A method for optional spacing is necessary. This could also address imbalances like in the "C#" example. (todo: create an issue)
Tajima: There can be problems with inserting a space using a space character. (details?)
There are two types of ambiguities: contextual (like in "C++", "T-shirt") and font-based (specific examples?).
bin-sensei: On the issue of imbalance - in Japanese, it's normal to have a space before "12" but not after "%" in the phrase "This is 12%". Therefore, imbalance does not necessarily mean bad.
Nat: We should come up with a proposal that better handles characters with mixed personalities, like those in the Unicode 2000 series. Some issues may not be resolved without distinctions like IVS. (todo: create an issue)
Nat: There's a need in CSS to locally define whether to space or not, based on character combinations or words. (todo: create an issue)
Taro: The desired amount of space varies with fonts in Western scripts too, depending on cap/x-height, and the font should include information about the space width between words.
However, it's difficult to apply the same for Japanese and Western character spacing since the ideal width varies depending on which Western font is matched with the Japanese font. The conclusion seems to be that CSS should allow specifying the size of the spacing.
Nat: When the space between Japanese and Western characters is set to 0, can it be used for the line-end alignment? What about when there is some spacing? bin-sensei: Use it for adjustment when it's 0; if there's 1/4em spacing, it should be the lowest priority.
Nat: It should be emphasized that the space between Japanese and Western characters is different from a word space. bin-sensei: Treat it the same as other Japanese character spacing when it's 0. There should be two patterns: one using it for character spacing and the other not. When used for character spacing, it should be spaced similarly.
From this discussion, it's evident that the issue is complex. It seems simply deciding a default width between character codes is not a sufficient solution.

Amount of Space Between Japanese and Western Characters

The discussion focused on the desirable amount of space, though no clear conclusion was reached. However, specific action items were identified for further consideration.

The width of space between Japanese characters and Latin characters in the direction of text compression [Wrong Japanese section title §3.3.2 #38 <https://github.com/w3c/jlreq/issues/38>]
The desired spacing varies depending on the font (size of Japanese character faces, spacing in Western text), size, and application. Should fonts be capable of handling this?
The JIS / JLReq specification of a quarter-width space is based on the conditions of the letterpress era. Apple uses 1/8th.
When the space between Japanese and Western characters is set to zero, spaces appear in justified alignment (?)
The nature of space between Japanese and Western characters is different from that of Western spaces (details needed).
Is it acceptable to use for adjustment? Yes, it is.
CSS should allow for adjusting the width.
For those entering Japanese typesetting from the web, it might be beneficial to explain the history of spacing between Japanese and Western characters on jlreq-d.
Using spaces other than the normal Unicode space for creating gaps between Japanese and Western characters can cause issues at line ends. Namely, they are not removed at the line end. This is correct behavior for a fixed-width specification (Mr. Murakami). However, there's a discussion about absorbing U+3000 at line ends. It's preferred to have it absorbed (Mr. Murakami, please provide the gh issue number).
todo: Send a document summarizing bin-sensei’s thoughts on spacing, etc., later (Bin-sensei).
Coordination with clreq would be necessary (c.f. https://www.w3.org/2023/11/17-clreq-minutes.html#t02).
Punctuation Spacing Processing

[css-text-4] text-spacing-trim and classes of closing punctuation [#9504]

Although there was no specific discussion on this, it is related to the spacing of consecutive punctuation and needs to be reviewed carefully. If the proposal here is agreeable, it should be reflected in jlreq-d.

Spacing of Japanese and Western styled text and punctuations

The above discussion defines spacing in terms of combinations of characters. However, in JLReq, the spacing is also influenced by how characters are used. Typically, this relates to whether glyphs are full-width or proportional. But it's difficult for text systems to distinguish whether a character's glyph is full-width. Furthermore, language settings often apply to the entire system, making it challenging to cater to users who use multiple languages, like Japanese people writing in English, where the same character may be preferred in both full-width and proportional forms. Perhaps the code should be separated. A few years ago, during a discussion on character classes, several problematic code points were identified. It might be worth organizing and discussing these at Unicode or a broader forum.

Nat: The discussion started with making small changes, initially handling quotation marks only with unicode IVS. Character class of quotation marks [#46 <https://github.com/w3c/jlreq/pull/46>]

The Dash Issue

Related to the above, there's confusion in Japanese about which Unicode character to use for the full-width dash: U+2014 vs U+2015. Can this be resolved?

Other Issues

Reviewed these issues and added comments as appropriate.

jlreq-d

Ways to Indicate Paragraphs [Fixed text in note for section 3.3.8 for Figure 139. #40 <https://github.com/w3c/jlreq/pull/40>]
Suggestion: How about writing that some people only want to read the first sentence of a paragraph?
Handling of consecutive brackets of different font sizes [Fixed figure 193 img3_1_9.png (former 4.9) #39 <https://github.com/w3c/jlreq/pull/39>]
Typesetting of tables in digital text [Fig 170 (img2_7_1) contains wrong chemical formula #36 <https://github.com/w3c/jlreq/issues/36>] (Does this need a conclusion?)
jlreq

todo: Update Ruby terminology wiki and solicit feedback from clreq [todo: Update Ruby terminology wiki and solicit feedback from clreq #385 <https://github.com/w3c/jlreq/issues/385>] → No harmonization with clreq
done: Are there better ways to manage comments and changes to draft level documents? [todo: Are there better ways to manage comments and changes to draft level documents? #384 <https://github.com/w3c/jlreq/issues/384>]
Translate new text in README.md [Translate new text in README.md #381 <https://github.com/w3c/jlreq/issues/381>]
clreq

Emphasis dots should be center-aligned with the text after the letter spacing is increased [#550] (JLReq Emphasis dots should be centre-aligned with the text after the letter spacing is increased #368 <https://github.com/w3c/jlreq/issues/368>)
Lack of support for applying extra spacing between Chinese/Japanese and Western text [#401]
Items with Consensus, Awaiting Reflection in jlreq-d

Develop a definition of 全角 that is generic enough to cover non-square typefaces [Fig 43 (img_ad1_12) English version has Japanese character in its description #35 <https://github.com/w3c/jlreq/issues/35>]
[css-values] Comments on the ic unit [#8769]
Annotations and Search [added note into text of section 2.6.1(c) #45 <https://github.com/w3c/jlreq/pull/45>]
Ideal spacing before fullwidth opening punctuations at the beginning of sentences or lines [character tables has omitted letters #37 <https://github.com/w3c/jlreq/issues/37>]
------------------------------------------------------------------------------------------
JLRreq TF 2023-11-24 F2F meeting notes

- Natが日本にいるのに合わせて数年ぶりのF2Fミーティングを行いました。

- 鈴木さんがフォントワークスから初顔合わせ。

- また、GitHub issue をベースに議論を行った最初のミーティングとなりました。まだまだやり方に改良が必要そうです。

出席者
木田、山口（大学院生！）、下農、山本、小林、田嶋、Nat、村田、小畠、敏先生、村上、鈴木（フォントワークス）

管理的な問題

現在Wikiで管理されているドラフトをレポジトリに移し、変更はGitHub issue & PRで処理する。レポジトリは ~/drafts の下。jlreq-d/sandbox/落書き帳.md に練習用のファイルを作成したのでそこで練習しよう。

todo: 木田: wikiにあるテキストをレポジトリに移す
todo: 下農: 編集方法のtutorialページをWikiに作る
その後、#384において、GitHub issueよりも気軽な方法として、GitHub discussions が良いのではないかとの情報。試し中。次回のミーティングで話す
内容に対してコンセンサスが得られてそれ以上の議論は必要ないが、まだドキュメントに反映されていない状態を示す方法が欲しい

和欧文字間クラスの再定義

CSSで再定義しようとしている。石井さんと私が、Unicode でプロパティとして定義する提案を準備中（12/12 [UTCに提出]）。

todo: 石井さんから提案のデータ内容をチェックし、差異ついて議論を行い（下のissues）フィードバックを行う。
todo: [spacing property]を約物のアキと漢字・欧文間の空きに分離し、また議論の結果を反映してアップデートする
issues
和欧間スペース：フットマーク用の記号 *†‡◊ [Combined and updated 3.3.8 f,g,h into one (3.3.8 f) #44 <https://github.com/w3c/jlreq/pull/44>]
和欧間スペース：絵文字 [span class=index in japanese text is used? #43 <https://github.com/w3c/jlreq/issues/43>]
和欧間スペース：組文字（㌀㎏）の扱い [fixed typo in section 3.3.8 note #42 <https://github.com/w3c/jlreq/pull/42>]
和欧間スペース：丸つき文字の扱い [fixed English version figure 119 (img2_3_14) #41 <https://github.com/w3c/jlreq/pull/41>]
[css-text] Extra spacing between ideographs and non-fullwidth punctuation/symbols [#9479]
日本の書籍で%の後はべた
[css-text][text-autospace] Is halfwidth Kana "non-ideographic letters"? [#9471] (JLreq [[css-text][text-autospace] Is halfwidth Kana "non-ideographic letters"? #378 <https://github.com/w3c/jlreq/issues/378>])
[css-text] The definition of ideographs includes punctuation marks [#9501]
[css-text] Add black square & white square characters to the definition of ideographs? [#9603]
議論

一つの文字でも場合によって、開けたい場合、開けたくない場合、両方の場合がある文字がある。どちらかに決めるとどちらかの用法で問題が残る。その場合に任意に調節する方法があるべき。空きを削除するよりも、空きを作る方が簡単に思われる。任意で開ける方法が必要。それをサポートできれば、”C#”の例のようにアンバランスが生じた場合にでも対処できる。（issue を作る）
Tajima: 文字でspacingを入れることにも問題があることもある
曖昧性には二種類ある。コンテキストによる場合と（C++、Tシャツ）、フォントによる場合（具体的に？）
敏先生: アンバランスになる問題について：「これは12%です」という文で12の前は開けるが%の後は開けない、というのは日本語では普通。なので、アンバランスが即悪いわけではないのでは？（敏先生）
Nat: mixed personallityの文字の扱いをどうにかして提案してほしい。unicodeの2000番台のようにIVSなどで区別が可能にならないと解決できないものもある（Nat）
Nat: 文字の組み合わせを指定して、開ける、開けない、をローカルに定義できる方法がCSSに欲しい。文字の組み合わせ、もしくは単語に対して定義。（issue作る）
Taro: フォントによって望ましいアキ量が異なる問題は欧文についても言えて、cap/x-heightにより違う、フォントの中に単語間空白の幅についても書かれているはず。
ただし、では和欧間で同じようにできるかというと、和字に対して、どのフォントの欧文を合わせるかによって望ましい幅が異なってくるので、困難。CSSで空き量のサイズを指定できるようにすべきだという結論だと認識。
Nat: 和欧間が0設定の時、行端揃えで調整には使っていいのか、感覚がある場合はどうするか。敏先生: 0の場合は調整には使う、1/4emが入っている場合は一番低い優先度にしてほしいと思われる。Nat: 和欧間はワードスペースとは違うものということを強調しておくべき。敏先生: 0の場合はほかの和文文字間と同じ扱いにする。文字間まで調整に使うときとそこまでは使わないときの2パターンで違うべき。文字間まで使うときは同じように開ける。
議論を見ていると、問題が山積。文字コード間のデフォルトの幅を決めるだけでは解決しない。

和欧文字間のアキ量

望ましいアキ量に関して議論を行った。特に明確な結論はないが、具体的な結論に落として行くべき項目も見られる。

和文文字とラテン文字の字詰め方向の字間の幅 [Wrong Japanese section title §3.3.2 #38 <https://github.com/w3c/jlreq/issues/38>]
フォント（和文の字面の大きさ、欧文の文字間）によって、サイズによって、用途によって望ましいアキ量が違う。フォントが持てるようにする？
JIS / JLReqが四分を指定しているのは活版時代の事情による。Appleは1/8
和欧間がゼロの場合に、両端揃えにすると、アキができる（？）
和欧間のスペースは欧文スペースとは性質が異なる（要詳細）
調整に使って良いか？使って良い
CSSで幅を調節できるようにすべき
webから日本語組版に入った人のために、jlreq-dで和欧文間の歴史？を説明すると良いのでは
和欧間の空きを作るのにUnicodeの通常のスペース以外を使うと行末の処理が問題。つまり行末に来ても削られない。固定幅指定なので動作自体は正しいと思われる（村上さん）。がU+3000は行末で吸収しようという議論がある。吸収してほしい（村上さん、gh issue 番号お願い）
todo: 空き量などについてまとめた文章を後で送る（敏先生）
clreq との調整が必要（c.f. https://www.w3.org/2023/11/17-clreq-minutes.html#t02 )
約物のアキ処理

[css-text-4] text-spacing-trim and classes of closing punctuation [#9504]

これに関して特に議論はなかったが、連続約物のアキに関する議論なのでちゃんと見ておく必要あり。ここの提案に同意できるならjlreq-dに反映。

スタイル付きの場合の和欧文字間や約物のアキの処理

上記の議論では字間を文字の組み合わせに対して規定している。しかし JLReq における字間は文字の使われ方にも影響を受ける。典型的にはグリフが全角かプロポーショナルか。しかし、テキストシステムが、その文字のグリフが全角であるかどうかを見分けるのは難しい。また、言語設定はシステム全体に適用されることが多いので、例えば日本人が英語も書くなどの場合は同じ文字に対して全角が望ましい場合と、プロポーショナルが望ましい場合が混在し、複数言語を使うユーザーを対象にしにくい。コードを分けるべきではないか。数年前文字クラスの議論を行った際に、問題となるコードポイントをいくつか拾い出した。整理してUnicodeやより広い場所で議論すべきかもしれない。

Nat: 最初は小さい変更で行こうということで、コーテーションマークのみunicode IVSで扱おうという議論が始まっている。コーテーションマークの文字クラス [#46 <https://github.com/w3c/jlreq/pull/46>]

ダッシュの問題

上記と関連するが、日本語において、どのUnicode文字を全角ダッシュに使うのかについて混乱がある。U+2014 vs U+2015。解決できる？

その他のissue

これらのissueをレビューして適宜コメントを付け加えた
###jlreq-d

段落の示し方 [Fixed text in note for section 3.3.8 for Figure 139. #40 <https://github.com/w3c/jlreq/pull/40>]
段落：一センテンス目だけ読んでゆきたい人がいますよ、と書くのはどう？
文字サイズなどが異なる括弧等が連続した場合の処理 [Fixed figure 193 img3_1_9.png (former 4.9) #39 <https://github.com/w3c/jlreq/pull/39>]
デジタルテキストにおける表の組版 [Fig 170 (img2_7_1) contains wrong chemical formula #36 <https://github.com/w3c/jlreq/issues/36>]（結論が必要？）
jlreq

todo: Update Ruby terminology wiki and solicit feedback from clreq [todo: Update Ruby terminology wiki and solicit feedback from clreq #385 <https://github.com/w3c/jlreq/issues/385>] → clreq とのハーモナイズは行わない
done: Are there better ways to manage comments and changes to draft level documents? [todo: Are there better ways to manage comments and changes to draft level documents? #384 <https://github.com/w3c/jlreq/issues/384>]
Translate new text in README.md [Translate new text in README.md #381 <https://github.com/w3c/jlreq/issues/381>]
clreq

Emphasis dots should be centre-aligned with the text after the letter spacing is increased [#550] (JLReq Emphasis dots should be centre-aligned with the text after the letter spacing is increased #368 <https://github.com/w3c/jlreq/issues/368>)
Lack of support for applying extra spacing between Chinese/Japanese and Western text [#401]
コンセンサスが得られているが、jlreq-dへの反映を待っているもの

√ Come up with a definition of 全角 that are generic enough to cover non-square typefaces [Fig 43 (img_ad1_12) English version has Japanese character in its description #35 <https://github.com/w3c/jlreq/issues/35>]
[css-values] Comments on the ic unit [#8769]
√ 割注と検索 [added note into text of section 2.6.1(c) #45 <https://github.com/w3c/jlreq/pull/45>]
√ Ideal spacing before fullwidth opening punctuations at the beginning of sentences or lines [character tables has omitted letters #37 <https://github.com/w3c/jlreq/issues/37>]