[todo] Research segment break transformation rules discussed in Level 3 and Firefox's rules, file an issue on csswg-drafts

[Yasuo Kida （木田泰夫）]

This is outcome from the F2F meeting on 2024-4-17.

---

4.3.3. Segment Break Transformation Rules

When English text is appropriately segmented, segment breaks are managed properly, typically treated as spaces between words. However, when Japanese text is segmented in the same way, segment breaks become spaces, inserting unwanted spaces between characters, which disrupt formatting and searchability. Text Level 4's description of 'line-break-transform' allows for the removal of segment breaks without turning them into spaces, depending on the context. However, implementing this requires advanced linguistic knowledge and a conducive environment for discussing complex cases, making practical implementation difficult. A strong call was made for CSS to provide guidelines.

Fantasai and Mr. Murakami mentioned that there were discussions about these rules in CSS3, but they were deferred due to difficulty in reaching consensus. A case for debate is when one side of a segment break is Kanji or Kana and the other side is Latin (mixed case). It was suggested that the following rules might be appropriate:

• If the full-width character is CJK punctuation: discard.
• If the language is set to Japanese, etc.: discard.
• Otherwise: keep.

As Firefox has been implementing a rule to remove such spaces for many years, this could be considered a baseline.

Like with text-autospace, a perfect rule is unattainable; rather, a baseline that most people can reasonably accept should be aimed for.

Actions

Research rules discussed in Level 3 and Firefox's rules, and Mr. Kida, along with Mr. Murakami, will create an issue on csswg-drafts to propose this.