Meeting minutes
Spacing between ideographs and non-fullwidth punctuation/symbols
xfq: I wrote a proposal in w3c/
xfq: there is also a proposal in Unicode related to this in https://
xfq: see also comment from the CJK & Unihan Group in #20 of https://
xfq: Note that in Japanese, it's normal to have extra spacing before "12" but not after "%" in the phrase "あああ12%あああ"
xfq: Personally, I think there should be extra spacing after the percent sign in Chinese
Eiso: If we bring this up to Unicode, they will ask what are the criteria for "normal" behaviour
Eiso: This cannot be based on the subjective opinion of some experts
Eiso: What happens if we add the extra spacing after the percent sign? Will there be any issue?
… I wonder what the rationale is to have extra spacing before European numeral but not after "%"
… There is no extra spacing to the right of the percent sign, but what if the left of the percent sign is a Kanji? Should we add extra spacing to the left of the percent sign?
… It's rare, but what should we do if it happens?
xfq: I think if they don't add extra spacing on the right, then they don't add extra spacing on the left either
Eiso: We need to consider the problem of classification, like "no extra spacing on both sides", or "no extra spacing on only one side"
Eric: The percent sign is also available in fullwidth and halfwidth versions
Eric: there should be no extra spacing for the fullwidth version
xfq: yes, and I think the jlreq folks think there should be no extra spacing for the halfwidth version either
Eric: OK
Eric: Personally, I don't agree that Chinese should use this convention
xfq: Bin-sensei said it's normal in Japanese
Eric: not for Chinese, I think
xfq: so different languages may have different behaviours
Eric: jlreq is based on JIS X 4051
xfq: any comment on my proposal in w3c/
[xfq introduces the proposal]
Eric: Before discussing specific code points, we need to discuss whether the general direction is correct
Eiso: the general direction looks good to me
… U+00B7 is also used in Western text
… If it is used in Western text, followed by Chinese characters, there should be extra spacing
… If it is used in Chinese text, then there should be no extra spacing
… the behaviour should be based on the language tag
Eric: yes, that's also the case for other code points shared by Chinese and Western text
Eiso: U+00B7 can also be handled by SVS like https://
Eiso: Some ethnic groups also use Chinese punctuation, but they use halfwidth glyphs instead of fullwidth glyphs
Eiso: like Miao people and Lisu people
Eiso: they can't use the guillemets as in French and Russian because the height is not enough
Eric: Mongolian also uses "《"
[Discuss punctuation marks and fonts]
Eiso: If most of our requirements are similar to Japan's, we can build on Japan's proposal and mainly write about the differences with them
Zhengyu: We should use language tags to distinguish behaviours
Eiso: In character classification, we need to separate out the different requirements of different languages
Zhengyu: yes
[Discuss 隐讳号 (CJK cross symbol), SVSes, and emoji]
Zhengyu: I am not very optimistic about the promotion of SVSes
Zhengyu: At least in the short term
Zhengyu: Language tags are easier
… SVSes are invisible and needs the support of input method
[Discuss the design of input method]
[Discuss the backwards deletion of SVSes]
Yijun: Adding language tags is also a burden
Zhengyu: SVSes are easier for professionals and harder for ordinary users
Yijun: but for user-generated content, it's not possible for them to add language tags
Yijun: if the input method supports it smoothly, SVSes should be transparent
Yijun: like emoji
https://
[Discuss SVSes vs language tags]
Eiso: I agree with Yijun
Eiso: SVSes should be transparent to users
Eiso: SVSes and language tags don't conflict. We can support both.
Zhengyu: agreed
Zhengyu: What I meant was we can't *just* rely on SVS
Eiso: moreover, the "%" problem cannot be solved by SVS, so language tag is required
xfq: also, we need to discuss the default behaviour first and then consider the exceptions
[Discuss currency symbols]
xfq: Personally, I think there should be extra spacing around currency symbols
Eiso: I think a lot of currency symbols are based on the Latin script, and it is reasonable to follow the same rule of Latin letters
Yijun: are there any characters that have extra spacing on one side, and no extra spacing on the other side?
Eiso: I don't think so
Zhengyu: "%" is a special case in Japanese. For Chinese, let's just think of it as Western text.
[Discuss the footnote marks]
[Discuss the example in https://
[Compare the example with https://
Eiso: if the rule is definded in a W3C document, it is easy to change
Eiso: but if it's defined in a Unicode character property, it's very difficult to change
Zhengyu: agreed
xfq: should we update our clreq document about this?
原则上,汉字与西文字母、数字间使用不多于四分之一个汉字宽的字距或空白。
Eric: we have "原则上", so maybe we don't need to mention all cases
… because there are many exceptions
… if we do want to add relevant text, maybe we can add a note
xfq: The higher priority problem is solving the CSS issue
… and then discuss whether we need to update clreq
Zhengyu: I think we should talk with CSSWG to make the behaviour be based on the language tag
Zhengyu: and then we start to talk about the edge cases
Eiso: Many ethnic groups in China and Vietnam still use Han characters, and the rules of use of Han characters are different from the normal Simplified and Traditional Chinese
Yijun: Can we distinguish by rules, not by language? For example, some people may prefer a predefined jis-x-4051 rule, even the language is Chinese.
Eiso: there is no rule in GB and CNS
Eiso: for example, if usage is relatively stable within the simplified characters of Mainland China, then the use of language tag is sufficient
Eiso: If the usage is different between Mainland China and Singapore, we can also distinguish by region
Eiso: I don't think that there will be more than one set of rules within Mainland China or Taiwan
Yijun: OK
Eiso: also for Japan, I think
Eiso: the current behaviour in CSS is not language-dependant, right?
xfq: correct
[Discuss language-dependant behaviour in CSS, like case transforms, text decoration, quotations, line breaking, hyphenation, justification, styling initials, etc.]
xfq: I can bring this up in CSS
Yijun: In addition to Western text in Chinese, we also need to consider Chinese in Western text
Zhengyu: In this case, most users will manually add spaces, but in some cases, for example, if a product name is a mix of English and Chinese, and the name appears in an English context, there may be a problem.
[Discuss edge cases]
Zhengyu: There's a lot more complexity.
Eiso: Let's put the update of the clreq document on hold for now and discuss it clearly with the CSSWG first
xfq: OK
Next teleconference time
February 21 (Wednesday), 19:00-20:00 (UTC+8)