Meeting minutes
Generic font family
Bobby: Taiwan, Hong Kong and Macao also use Fangsong
… Taiwan's official documents don't use Fangsong, but Fangsong is common in books
… however, macOS only has a SC Fangsong font
Eric: It's a chicken and egg problem, Taiwan/HK/Macao government don't require Fangsong, so the operating systems don't provide TC Fangsong
Eiso: Macao has a relatively short history of publishing, with relatively few publishing practices, compared with Hong Kong
Eiso: there's another problem, if we remove the cursive = kai equivalence, what will cursive map to in a zh environment?
xfq: As I understand it, it won't map to anything
Eiso: maybe 行書 (semi-cursive)?
https://
Eric: or 草書 (cursive)
https://
[Discuss generic font family for Western typefaces]
Eric: We haven't even solved the basic typographic features of Chinese yet, and it's too early to discuss more advanced typography
Bobby: If we need to wait ten years for the browsers to implement Kai and Fangsong, typographical needs in Chinese will never be met
Bobby: the CSSWG and the browser vendors need to implement Kai and Fangsong quickly
Eric: agreed
Eiso: +1
xfq: +1
Zhengyu: We can't expect browser vendors to know all the major Fangsong fonts
… Safari supports ui-monospace, but Chrome does not
Zhengyu: developers want ui-monospace badly
… and it's not possible to use "font-family: 'San Francisco'" because Apple does not publicly expose the font
… we need both browser vendors and operating systems support to make it work
[Discuss the system fonts in Apple operating systems]
Zhengyu: regarding Kai and Fangsong, until the browsers and operating systems support it, we can encourage developers to use Web fonts first
Zhengyu: regarding https://
Zhengyu: for example, Fangsong fonts can also have Latin glyphs, which contain Latin letters of similar style to the Chinese Fangsong font.
Zhengyu: If you make a font that conforms to the Chinese national standard, there has to be Latin letters in it
Eiso: what is the style of the Latin letters in a Fangsong font?
Zhengyu: This is something for type designers and the industry to consider
Zhengyu: But it's possible to include Western letters in a Fangsong font
Zhengyu: It's a design issue, not a technical issue
Eiso: This is also influenced by region.
Eiso: Latin letters in Kai fonts in China are serif, while Latin letters in North Korea's Kai fonts are more similar to the traditional Kai style
Zhengyu: We can't ask people to make a particular style of Western letter glyphs in a particular font style
Zhengyu: Because even different Western serif fonts can vary greatly from one another
xfq: But there are many fonts that don't have Latin letters in them
xfq: like many Arabic fonts
Bobby: Yes, but there are also a lot of fonts that do have Latin glyphs in them
Spacing between ideographs and non-fullwidth punctuation/symbols
[xfq introduces the issue]
xfq: we're discussing this in CSSWG
… Chromium has implemented text-autospace, but it's not in stable channel yet
… any comments?
Eric: Unlike jlreq and klreq, we don't categorise characters
https://
https://
[Discuss the character classes in jlreq and klreq]
xfq: We can consider mentioning the punctuation/symbols that we think need to have extra spacing first
xfq: and for the ambiguous ones and the ones that don't need extra spacing, we can discuss them later
xfq: maybe start from code points in ASCII
[Discuss code points in ASCII]
Eric: U+002F SOLIDUS [/] has no extra spacing around it according to the national standard, but some people think it's too tight and add spaces. I think either is fine.
Eric: we can define things like "Extra Spacing", "Set Solid", and "Ambiguous"
Eiso: is the scope all characters that have been encoded?
Eiso: Or do we only consider all the characters used in the Mainland China, Hong Kong, Macao, and Taiwan standards, as well as characters that are not in those standards but are actually used?
… We need a starting point
Eric: We need to consider at least all the punctuation marks in clreq, and all the code points in ASCII
[Discuss the degree Celsius symbol]
Zhengyu: I think the degree Celsius symbol can be treated the same way as ideographs
… I don't think it matters if there's no extra spacing to the right of U+2103
… I agree with hax's comment in w3c/
… we can't expect the default behaviour to solve all microtypography problems
… The list of exceptions is never-ending, and in this case I don't see the problem
[Discuss the behaviour in Chromium]
Bobby: I think we can add extra spacing to all non-fullwidth symbols by default.
xfq: including emoji?
Zhengyu: yes
[Discuss the emoji case]
https://
[Discuss East Asian Width constants in ICU]
[Discuss whether halfwidth Kana is considered "non-ideographic letters"]
Eiso: there are also halfwidth versions of hangul
Eric: emoji presentation sequences behave as though they were East Asian Wide, regardless of their assigned East_Asian_Width property value.
xfq: about Nüshu, the width of the character is less than the height of the character
Eiso: East_Asian_Width property value of Nüshu characters are East Asian Wide
Eiso: To be honest, this kind of thin Nüshu font style only appeared in the 90s, it used to be thin, but not this thin
xfq: Let me put together an initial proposal that has three categories: Extra Spacing, No Extra Spacing, and Ambiguous
Eric: refer to East_Asian_Width property value of the code point by default, and only list the code points that are inconsistent with East_Asian_Width
Eiso: Extra Spacing for H/N/Na, No Extra Spacing for F/W, Ambiguous for A by default
Eiso: U+25A0 and U+25A1 are A, but they can be used to represent a missing ideograph, and must be treated in the same way as Han characters in Chinese contexts
Eiso: It doesn't affect ordinary texts much, but for poetry, it's likely to be misaligned
Eiso: there was a proposal to add separate code points for black square & white square characters used in ideographic text, but it was refused by the UTC
Zhengyu: we can solve this by adding span elements
… Is this feature on by default?
xfq: yes
… see w3c/
Zhengyu: I think this feature should be turned off by default
Zhengyu: because it's a microtypography feature
… A lot of people don't even notice it
Eric: yes, also for text-spacing-trim
… text-spacing-trim sometimes has a negative effect
… although we could turn it off
xfq: it's on by default in Microsoft Word, iOS UI, and WeChat
Zhengyu: Word processors may be a good example, but iOS is by no means a good example of microtypography, as we found some problems
… We should look into common examples rather than special cases
Eric: Our document should solve the most basic problems first, and then move on to other more advanced problems, like Chinese opera books
Eiso: agreed
w3c/clreq#109
Bobby: single-line inline annotation can be implemented by using font-size and vertical-align
Eric: we need to document the requirements first
[Discuss the example in w3c/
Eiso: when doing string searching operations, should the annotation be skipped or included, this is a question
Eiso: in the first image in w3c/
Eiso: For longer explanatory text (either single-line or double-line), the annotation probably should be skipped
[Discuss various comments in w3c/
Eiso: I have no comments on the draft proposal in w3c/
Eiso: but we need more pictures
Eric: can you translate the text in w3c/
huijing: sure
xfq: I made some comments on w3c/
Eric: I'll look into the comments
Eric: I agree with not using the word warichū
[Discuss the English terminology]
[Discuss warichū embedded inside a warichū]
Eiso: there are real examples of warichū embedded inside a warichū
Eric: We need illustrations about the reading order
… about English translation, we can use "inline annotation"
Eiso: agreed
xfq: so we use "single-line inline annotation" and "double-line inline annotation"?
Eric: agreed
Eiso: agreed
Next teleconference time
December 15 (Friday), 19:00-20:00 (UTC+8)