[css-values] Comments on the ic unit

https://www.w3.org/TR/css-values-4/#ic

We discussed the ic unit at yesterday's Chinese Layout Task Force Teleconference. We found two issues about the current spec text:

First, the wording 'CJK letters' in the description is not accurate. We know that it doesn't affect the normative definition of the ic unit (which is the advance measure of U+6C34), but we want the wording used here to be more accurate, because 'CJK letters' can be understood as one or several of:

• Han characters
• Hiragana
• Katakana
• Small kana
• Iteration marks
• Prolonged sound mark in Japanese
• CJK Strokes
• Chosŏn'gŭl/Hangul jamos and syllables
• Maybe even Bopomofo and Nüshu

Second, we want the ic unit to have a fallback mechanism, since the water character (U+6C34) might not be included in the font subset. We noticed this sentence:

In the cases where it is impossible or impractical to determine the ideographic advance measure, it must be assumed to be 1em.

but it doesn't explicitly state that this is a fallback mechanism, which we would like to clarify.

In addition, Hangul also has fixed width fonts. We already have ch and ic, but Korean may need a similar unit. This may need to be discussed with the klreq group.

Regarding the choice of the character (U+6C34 or something else), we have no strong preference, but if the CSSWG would like to choose it based on the character frequency, there are more common characters than the water character in both Chinese and Japanese.

PS: Chinese, Japanese, and Korean do have some proportional width fonts, including condensed type, expanded type, and other kind of proportional width fonts. That said, we don't think people using this type of font need to use the ic unit, so we think it's fine.

Start by reading the CSS Values and Units Level 4 section linked in the issue and the surrounding definition of the ic unit. The issue requests more precise wording and an explicit fallback, while the possible Korean unit and reference character require CSSWG and klreq discussion; completion depends on resolving those questions before editing the spec.