Meeting minutes
Adjustment of adjacent punctuation marks
https://
xfq: CSS has adjacent-pairs trimming ^
https://
"In principle, any two adjacent punctuation marks that are 2 em wide should be reduced to 1.5 em. Following the same principle, they may be reduced to 1 em for certain typographic styles."
xfq: in clreq, we allow two adjacent punctuation marks to be reduced to 1 em for certain typographic styles ^
xfq: This is currently not supported in CSS. Do you think it is necessary to support it?
Eric: This also involves prohibition rules for line start and line end and full justification, and the issue of who takes precedence
[The group reviews CSS Text L4]
[Discuss the differences between print and the web]
Eric: The line length on the web is often uncontrollable
… I think this matter can be put aside for now because it is very complicated
… and it's currently only supported by Blink
Bobby: WebKit is likely to have the same behavior as iOS, and may not necessarily develop a new set of behaviors for the web.
Eric: but the behavior on iOS is not correct
Bobby: indeed
Eric: and fonts must have the OpenType `halt` or `chws`
Eric: Most fonts don’t have `chws`
xfq: indeed
Eric: `chws` is a relatively new feature
Eric: not in the initial version of Source Han Sans Chinese, for example
… supported by Hiragino, PingFang, and Source Han Sans
… not supported by Microsoft YaHei
[Discuss the current state of system fonts]
Eiso: Many font developers in Mainland China don’t know this feature, and they don’t even know many old OpenType features.
Bobby: same in Taiwan
Eiso: Mainland China, Hong Kong, Macau, and Taiwan may need to have local official guidance on fonts.
[Discuss the typeface companies in Mainland China and Taiwan]
Eric: Japanese fonts have a de facto standard, but Chinese font do not
… Chinese font developers usually only support the horizontal and vertical glyph conversion feature
Eiso: Many fonts don’t even support this, and these fonts are used in vertical writing mode. The rendered results are funny.
Bobby: Is it because font developers don’t know these features, so the new fonts don’t support them, or are too many people still using old fonts?
Eiso: Both.
Zhengyu: this is a chicken and egg problem
… First we need to make the system fonts support it
… like the Microsoft fonts
Eric: We should build powerful layout engines so that no matter how stupid the font is, they can render the glyphs correctly.
… Should we ask the people responsible for fonts at Microsoft to support halt/chws as a group or as individuals?
… they can update the Windows fonts
Bobby: who is in charge of maintaining Chinese fonts on Windows?
Eric: I don't know, they may need to ask for outside help
… it might be more difficult to make fonts from the Founder Group to support halt/chws
… but we can deal with the OS fonts first
… There are also problems with some Japanese fonts
… for example, the punctuation marks in Yu Gothic UI are incorrect
… Yu Gothic is correct
… because Microsoft modified the Yu Gothic font incorrectly
… fonts from Google and Apple have similar issues
[Discuss the punctuation width adjustment convention in Taiwan]
Yijun: in Taiwan, the width of periods should not be adjusted
Eric: in Simplified Chinese, the width of [。”] should be adjusted
xfq: The period can be seen as a hollow middle dot
… according to https://
Eric: 1/4 em before and after the character face
[Discuss how to do the punctuation width adjustment]
https://
xfq: requirements for Japanese ^
xfq: see also https://
Eric: This is different between Chinese and Japanese
xfq: in https://
Yijun: I think the centered comma, period, secondary comma, colon, semicolon and interpuncts in https://
Eric: Some people who don’t understand Chinese think this is kerning, but it’s not.
Eric: Because kerning is static
Eric: but this is dynamic
Eric: There is a big difference between the current CSS spec and real mojikumi. The advanced mojikumi algorithm is much more complex.
https://
[Discuss the glyph of interpuncts in Simplified Chinese]
"The width of interpuncts varies in different regions. In principle, in Hong Kong and Taiwan, interpuncts should have the same dimensions as a Hanzi in both vertical writing mode and horizontal writing mode. In Mainland China, interpuncts take up half the space of a Hanzi."
Eric: 'text-spacing-trim: trim-all' can be used in some Chinese dictionaries
Yijun: indeed, and this is not kerning
[Discussing kerning in Chinese characters, Kana, and Hangul / Chosŏn'gŭl]
https://
Eric: there's kerning in フィー, for example
Bobby: I personally think there is no demand for adjustment of adjacent punctuation marks in Taiwan
Zhengyu: What are the publishing practices in Taiwan?
Bobby: no adjustment
Zhengyu: OK
Bobby: except the compression of punctuation marks at line start or line end
Zhengyu: There is another issue. I think when lang is a non-CJK language, text-spacing-trim should be space-all.
xfq: I'll create an issue to discuss https://
Go through the pull request list
https://
w3c/clreq#596
All: OK to merge
w3c/clreq#603
[xfq introduces the discussions in the i18n WG meeting]
xfq: I also added a link to the Contributors section in the spec metadata section
Eric: looks good to me
Bobby: looks good to me
xfq: I'll update the PR
Next teleconference time
April 24 (Wednesday), 19:00-20:00 (UTC+8)