You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
A trial of posting a meeting notes as a GitHub issue. The expectation is that it can capture follow up discussions in one place. Also, issues stem from the discussion or related issues can easily be inter-linked to give them a context.
Shapes of some characters in vertical orientation are inconsistent depending on the combination of the font and the application used. Such inconsistencies are problematic especially with ebooks. Solving the issue in open environment where you can’t assume specific combinations require a good open standard that fonts can agree and follow. Similar issue exists in controlling spacing around Japanese punctuations. It requires cooperation from fonts as applications do not know where and how much extra space exists in a glyph. To help solving the issue it was agreed to create a font specific discussion thread in JLReq TF.
The first meeting on 3/30 discussed issues Murata-san raised regarding vertical orientation shapes. The second meeting on 4/20 discussed a proposal of ‘chws’ OpenType font table which allows shaping engines to adjust spacing between punctuations in a good performance. Font experts from CITPC including font venders joined from this meeting. This meeting will continue discussing punctuation spacing especially cases that are not covered by ‘chws’ table.
Today’s goal
‘chws’ OT table does not cover space adjustment at the beginning and ending of a line, and also when it spans multiple style runs. Understanding options for supporting them.
Are ‘halt’ and ‘chws’ provide complete solutions for various applications to appropriately support punctuation spacing?
With the previous meeting we understood that adding new features to fonts is a big change / investment for font vendors. Therefore, we would need to propose a complete set of features next generation fonts would support instead of requesting them to implement individual features in piecemeal fashion. In this regard we want to understand how cases of punctuation spacing that are not covered by ‘chws’ will be supported. esp. if ‘halt’ and ‘chws’ are everything we want them to support in terms of punctuation spacing.
At the same time like mentioned by Tajima-san, there are areas such as EPUB readers, OS standard text and web font venders that can move more quickly.
Brief introduction of low level software that implements layout (Ishii)
Shaping engine is the low level framework for text layout. Effectively there are only three of them.
Halfbuzz by Google, used by Androids, Chrome, Firefox and Linux
CoreText by Apple used by iPhones, iPads and Macs
DirectWrite by Microsoft used by Windows
Halfbuzz process one style run at a time. It is closer to OpenType library
CoreText and DirectWrite can process multiple runs / lines at a time
Because of the difference sometimes the same feature needs to be implemented in a different layer depending on the shaping engine used.
Two approaches for punctuation spacing at the beginning / ending of a line (Ishii)
One approach is that applications apply ‘halt’ to the character at the beginning and ending of a line. This is similar to how optical bounds are supported for Latin typeface. It only require ‘halt’ from the font. As it is applied to only two characters in a line, the impact to the performance would be at a reasonable level.
The other approach is to insert a dummy character before and/or after a line before giving it to the shaping engine so ‘chws’ process them as if the beginning and ending of a line is a character.
Further discussions would be required to determine which of the two approaches should be used (perhaps the answer is different between shaping engines?)
It is harder to solve the cross style run case (Ishii)
It would require three steps - processing the preceding style run to trim the extra space at the end of the run, processing the following style run to trim the extra space at the beginning of the run, inserting some space in-between.
However…
for the first place it is not clear how much space is necessary when styles are different between adjacent punctuations (Ishii)
Cases like different font sizes, when the font is proportional such as So-sho (brush stroke), etc.
When it is within a style run, ‘chws’ gives designers a freedom to change the amount of space. It is an advantage of ‘chws’ approach (Ishii)
Right now the space is always half-width but it came from the limitation of metal type setting. (Bin-sensei)
JLReq does not mention how such cases should be processed. However X4051 touches on it it is not useful as it assumes specific cases that do not apply any more (Bin-sensei)
There would be two ways, one is to take an average of a half width of two sizes. The other is to apply a half width of the bigger one. I think the latter is more preferable. (Bin-sensei)
Shouldn’t we add a note to JLReq? Could you (Shimono-san) create a GitHub issue? thanks. (Kida)
Will write up something for the multiple font size case and send it out to the mailing list (Bin-sensei) (done)
It would require completely different thinking / system for proportional Japanese brush stroke styles such as gyo-sho (Nat)
The half-width space is used not necessarily because it creates the better layout. There has been discussions that a half-width space is too wide. Quarter-width corner bracket existed. etc. (Bin-sensei)
The half-width-ness was a requirement from efficiency. Characters that are not in unit width create extra space at end of lines and such lines required manual adjustments = extra work in fully justified style. (Bin-sensei)
Well, it seems the discussion is coming to the logical conclusion that Japanese text should adopt rugged-right layout :) (kida)
People are not accustomed to it but I think someday it will be accepted. European languages were also fully justified but nowadays rugged-right is far more popular (Bin-sensei)
Bin-sensei reviewed ’chws’ layout samples. How’s the result?
Commented on how quotation marks are rendered and sent it out to the list.
Conclusion: to support punctuation spacing are ‘halt’ and ‘chws’ the only tables font should prepare? (Kida)
It is possible that advanced layout software need some more (Ishii)
‘chws’ can’t be used with other features (Nat) Could you elaborate? (Kida)
What is the best way to communicate with people joining from CITPC? (Kida)
JLReq TF list is not limited to font related issues. But if you are interested you are more than welcome to join. Please contact me or Shimono-san.
Another possibility is JLReq GitHub Issues. (However it would not work well for administrative communications such as meeting reminder). We can also use CITPC GitHub (the same issue) or MS Teams (not good at archiving).
There might be people who want to join JLReq TF. There are cases, however, the background of discussion is not clear (to new people). (Tahara)
Will try to cover background as much as possible (kida)
next meeting
The next meeting is on 6/15 10 am JST. Will come back to the vertical orientation issue based on reports Nat-san will share.
The text was updated successfully, but these errors were encountered:
One reason that rugged right has not become popular in Japanese layout might be that because most characters are multiple of full-width naturally the length of most lines become consistent, with small number of lines in a different length. This is unlike English case where all lines are equally uneven. When only a small number of lines are shorter it looks as if it has a special meaning (e.g. it might be an end of a paragraph). also it does not look good.
I think in case of Japanese rugged right needs to be combined with folding lines at word or phrase boundary. Folding lines at word or phrase boundary is only possible with rugged right (because adjustments becomes too large).
When automatic word / phrase segmentation becomes easily available I think rugged right will also be accepted.
A trial of posting a meeting notes as a GitHub issue. The expectation is that it can capture follow up discussions in one place. Also, issues stem from the discussion or related issues can easily be inter-linked to give them a context.
Previous meeting: #269
–––––––––––––––––––––––––––––––––––––––––
(English translation follows)
JLReq TF Meeting notes draft フォント分科会 5/18
背景
文字の縦書き字形には実装による不統一があり、電子書籍の作成などで課題になっている。また、約物間の半角詰めをOSレベルで実現するためのテーブルが提案されている。これらの機能の実現にはフォントとレイアウトが協調して動く必要があり、つまりオープンな仕様が必要である。フォントとレイアウトエンジンとの間の問題点を理解するため、フォントに知識のある人を招いてフォント分科会を作った。
第一回 3/30 のミーティングで村田さんの提起した縦書き字形の問題を議論した。第二回 4/20 のミーティングでは CITPC からのメンバーが加わり、石井さんの提案による約物間の半角詰めをOSレベルで実現するための OpenType テーブル chws について議論した。今回のミーティングでは、chws テーブルでカバーされない約物半角詰めについて議論する
今日のゴール
chws テーブルでカバーされない約物半角詰めをサポートする方法のオプション、フォントに何が必要かを理解する
参加者
CITPC
W3C / JLReq
議論
イントロダクション(木田)
前提として Shaping engine の説明(石井)
行頭行末は二つの可能なアプローチがある(石井)
マルチフォントはもっと厄介(石井)
そもそもどうすべきかが不明確(石井)
chws でのレイアウトのレビュー結果(敏先生)
結論として、約物詰めのためにはフォントに chws と halt があれば ok(木田)
文字情報技術促進協議会から参加してくださっている方々との議論・コミュニケーションの方法(木田)
次回
––––––––––––––––––––––––––––––––––––––––––––––––––––
JLReq TF meeting notes, font sig 4/20
Background
Shapes of some characters in vertical orientation are inconsistent depending on the combination of the font and the application used. Such inconsistencies are problematic especially with ebooks. Solving the issue in open environment where you can’t assume specific combinations require a good open standard that fonts can agree and follow. Similar issue exists in controlling spacing around Japanese punctuations. It requires cooperation from fonts as applications do not know where and how much extra space exists in a glyph. To help solving the issue it was agreed to create a font specific discussion thread in JLReq TF.
The first meeting on 3/30 discussed issues Murata-san raised regarding vertical orientation shapes. The second meeting on 4/20 discussed a proposal of ‘chws’ OpenType font table which allows shaping engines to adjust spacing between punctuations in a good performance. Font experts from CITPC including font venders joined from this meeting. This meeting will continue discussing punctuation spacing especially cases that are not covered by ‘chws’ table.
Today’s goal
Attendees
CITPC: Tahara-san (Topan), Nakauchi-san (Morisawa), Mizuno-san (Iwata), Kyo-san, Miyata-san (Dainippon Printing)
W3C / JLReq: Ishii-san, Kobayashi-san, Shimono-san, Tajima-san, Nat-san, Bin-sensei, Murakami-san, Murata-san & Kida-san
Discussions
Introduction (kida)
Brief introduction of low level software that implements layout (Ishii)
Two approaches for punctuation spacing at the beginning / ending of a line (Ishii)
It is harder to solve the cross style run case (Ishii)
for the first place it is not clear how much space is necessary when styles are different between adjacent punctuations (Ishii)
Bin-sensei reviewed ’chws’ layout samples. How’s the result?
Conclusion: to support punctuation spacing are ‘halt’ and ‘chws’ the only tables font should prepare? (Kida)
What is the best way to communicate with people joining from CITPC? (Kida)
next meeting
The text was updated successfully, but these errors were encountered: