Regional variant characters should be marked

[dahlia]

Sometimes, an existing character can be a regional variant of another character due to its coincidental reading in a region. However, often such regional variants cannot convey the intention in other regions.

For example, 才 (talent) is used often instead of 歳/歲 (age) in Japan due to its coincidental reading in Japanese: さい (sai). However, those two characters don't share the same reading in other regions, e.g., 才 is 재 (jae) & 歲 is 세 (se) in Korean.

If such regional variants replaced orthodox characters in a regional standard they should be normalized to orthodox ones in keys to group cognates, whereas their readings from other languages should be read as variant characters which aren't normalized.

For example, since 函 (box) is unlisted in Japanese jōyō kanji (常用漢字), 函 (かん; kan) is replaced by 関/關 (relation) which has the same reading. That's why 函數 (function) is written as 関数/關數 in Japanese. Still, as those two characters are read differently in other regions (e.g., 함 ham vs 관 guan in Korean), such uses need to be marked in the data set and pronunciations in <ruby> should show the proper reading (e.g., 함 instead of 관).

See also: #9.