Re: [csswg-drafts] [css-text] Use the Unicode East Asian Auto Spacing for `text-autospace` (#11013)

The CSS Working Group just discussed ``[css-text] Use the Unicode East Asian Auto Spacing for `text-autospace` ``, and agreed to the following:

* `RESOLVED: Adopt forthcoming UTR#59 as basis for text-autospace character classifications (deferring question of default behavior to a new issue)`

<details><summary>The full IRC log of that discussion</summary>
&lt;chrishtr> koji: this is about text autospace<br>
&lt;chrishtr> koji: there have been many discussions about this, now I'd like to bring it to the group<br>
&lt;chrishtr> koji: previous discussions led to complications<br>
&lt;chrishtr> koji: I and two other people made a proposal related to this to the Unicode working group, and it was accepted<br>
&lt;chrishtr> koji: it will be in public review soon as part of version 59<br>
&lt;chrishtr> koji: proposal is to switch the character categorization for autospace to align with version 59 of Unicode<br>
&lt;florian> q+<br>
&lt;chrishtr> rossen: and how long will version 59 take to publish?<br>
&lt;chrishtr> koji: not totally sure, maybe Jan or Feb?<br>
&lt;chrishtr> florian: as you said the set of issues were complicated<br>
&lt;Rossen1> ack florian<br>
&lt;chrishtr> florian: in principle, I think it's great that it's been taken up by Unicode and we should use it<br>
&lt;chrishtr> florian: follow-up questions. Some of the things we discussed previously were not regular full-width characters but some special cases, and how to deal with those. Am I right that this Unicode proposal deals with all those special cases in CJK?<br>
&lt;chrishtr> florian: there were also special cases about symbols, are those also handled?<br>
&lt;florian> s/some special cases/hangul and bopomofo<br>
&lt;chrishtr> koji: similar to utr 50, this doesn't do exactly the same thing as e.g. MS Word as it relates to these half-width or ambiguous characters, but it is close.<br>
&lt;chrishtr> koji: there are cases where people will want to override to adjust behavior, for things like printed typography<br>
&lt;chrishtr> koji: what we wanted to do in version 59 is to set a solid default value for code points<br>
&lt;koji> The current draft: https://www.unicode.org/L2/L2024/24259r-tr59-1-working-draft.html<br>
&lt;chrishtr> florian: the original thing had special cases around letters and numerals, but also possibly around symbols<br>
&lt;chrishtr> koji: the current draft of the Unicode proposal (posted in chat) classifies values to W and N and if CSS wants more control we will need to use general categories to split W into multiple pieces.<br>
&lt;chrishtr> koji: I'm not sure if we want to split symbols and punctuation<br>
&lt;chrishtr> florian: I'm not sure we disagree, just noting that we had discussion<br>
&lt;fantasai> https://www.unicode.org/L2/L2024/24259r-tr59-1-working-draft.html#symbols-punctuation<br>
&lt;chrishtr> florian: I'm supportive of the intent of your proposal, just haven't had the time to read it<br>
&lt;chrishtr> rossen: does this mean you're ok accepting this as a forward-looking resolution?<br>
&lt;chrishtr> florian: depends on if others have reviewed, might suggest changes on review<br>
&lt;chrishtr> fantasai: suggest accepting this change since it's an improvement, and if we find adjustments we can modify the Unicode proposal<br>
&lt;chrishtr> koji: just wanted to note that the proposal in Unicode was reviewed and accepted already, as an indication of review<br>
&lt;chrishtr> jfkthame: noticed that it's proposed that the changes are disabled outside of CJK for now, is that right?<br>
&lt;chrishtr> koji: yes<br>
&lt;chrishtr> jfkthame: what about content that is Chinese but not marked as such (e.g. blog comments)<br>
&lt;chrishtr> koji: this is a good point. the reasons I proposed it that way is that the property value is different between Chines and non-Chinese. If we applied to a document that is not Chinese we need a default for that document.<br>
&lt;chrishtr> koji: there's a chance that it would do the wrong thing on such documents<br>
&lt;chrishtr> koji: another reason is performance. This will slow down layout 1-3% due to additional complexity, and so limit it to the languages where it has clear user benefit.<br>
&lt;chrishtr> jfkthame: reasonable answers, but not sure about the conclusion<br>
&lt;chrishtr> koji: I think this is in line with what we've done for other properties<br>
&lt;chrishtr> jfkthame: agree that it's aligned with things like hyphenation. But in the case of hyphenation it could be wrong, whereas here it wouldn't be wrong but it would be a bit less good than it might be.<br>
&lt;chrishtr> florian: things will improve less for Japanese and Korean than Chinese?<br>
&lt;chrishtr> koji: it wouldn't solve the performance issue to include non-explicit languages<br>
&lt;chrishtr> koji: sub-optimal rendering would also result<br>
&lt;Rossen1> ack fantasai<br>
&lt;chrishtr> fantasai: wanted to express the same thing as Jonathan Kew<br>
&lt;chrishtr> fantasai: propose to accept UTR 59 but leave the question of in what situations to apply the behavior open<br>
&lt;chrishtr> fantasai: I lean towards what Jonathan and Florian are saying and trying to make it work when we can<br>
&lt;chrishtr> koji: I can file a new issue. To clarify, we're leaving undefined if the language is not specified, but not do it in English document?<br>
&lt;chrishtr> fantasai: I think that's open. We should ideally try to make it work in all documents if we can find a way.<br>
&lt;chrishtr> florian: agree with the proposed resolution<br>
&lt;fantasai> PROPOSED: Adopt forthcoming UTR#59 as basis for text-autospace character classifications (deferring question of default behavior to a new issue)<br>
&lt;florian> s/the proposed resolution/the proposed two-part resolution<br>
&lt;chrishtr> RESOLVED: Adopt forthcoming UTR#59 as basis for text-autospace character classifications (deferring question of default behavior to a new issue)<br>
</details>


-- 
GitHub Notification of comment by css-meeting-bot
Please view or discuss this issue at https://github.com/w3c/csswg-drafts/issues/11013#issuecomment-2504430075 using your GitHub account


-- 
Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config

Received on Wednesday, 27 November 2024 17:31:38 UTC