w3c
Baseline
We should mention that Chinese characters have no baseline. Each glyph is located in a square character frames of equal dimensions, with no ascenders or descenders.
我们需要提一下汉字没有基线，每个字形都位于正方形的文字外框中，既没有升部也没有降部。
We should mention how to deal with Chinese characters when mixed with scripts that has a baseline.
需要提一下，与有基线的文字混排时应如何处理。
We need to be careful about what 'baseline' means. CSS and other sources say that Han characters do have a baseline, but it's an ideographic baseline (ideo) that runs below the Han character glyphs, and is somewhat lower than the English baseline (romn). There are other types of baseline that that found in English text; northern Indian baselines are often said to be hanging and align with the top bar in certain scripts (hang). These baselines are used to relatively position the text in differing fonts in the vertical axis.
What is true is that Han characters have no descenders.
Some related issues:
- https://github.com/w3c/jlreq/issues/246
- https://github.com/w3c/csswg-drafts/issues/5381
The creation of the "ideographic" baseline is in my opinion a stop-gap measure, that lacks a full appreciation of how ideographic languages are typeset. When children learn to write them they use boxes and not lines for a reason -- that reason relates to how the characters occupy a space and can be written either horizontally or vertically. More important to the "baseline or not" argument is how they are arranged on the line with different fonts and different sizes -- there is no single choice but rather a number of alignment points to consider, and when set next to a language that is baseline-based, how that language's baseline should relate to the boxes involves a deep discussion. Each writing script that doesn't use a single baseline will fall into this category of "needing to be harmonized with writing scripts that use a baseline", and those other scripts should not be misunderstood to have a "baseline" that is just on some other y-offset from the Latin baseline.
