source-han-sans
source-han-sans copied to clipboard
The scale of some dingbat glyphs is not consistent across regions
The "Dingbats" font in Source Han Sans contains a number of region-specific glyphs for enclosed CJK characters:
This issue is about the scale the glyphs for other regions isn't consistent with those for JP. As you see in the picture below, the JP version looks fine, but the glyphs of ㈣, ㈥, ㈦, ㈧ for HK are a tad smaller than the others:
- BTW, I also wonder why the parenthesis of ㈩ is different from the others.
The scale of the JP glyph matches the non-JP ones when I resize the JP glyphs to 98%, so it seems that the non-JP glyphs had (inadvertently?) gone through some post-processing routines as the ideographs.
Note: Looks like it only affects the characters in the Enclosed CJK Letters and Months block (3200–32FF). Those in the Enclosed Ideographic Supplement (1F200–1F2FF) block are not affected:
For what it's worth, this is not a regression in v2.003. But the glyphs of TC/SC regions in v1.004 are fine (not scaled to 98%) though.
Below are the glyphs with incorrect scalings:
U+3223 ㈣: uni3223-CN
U+3225 ㈥: uni3225-CN
U+3226 ㈦: uni3226-TW
U+3227 ㈧: uni3227-CN, uni3227-TW
U+322B ㈫: uni322B-TW
U+322E ㈮: uni322E-CN
U+3232 ㈲: uni3232-TW
U+3233 ㈳: uni3233-CN
U+3235 ㈵: uni3235-TW
U+3237 ㈷: uni3237-CN
U+323C ㈼: uni323C-CN
U+323D ㈽: uni323D-CN
U+323E ㈾: uni323E-TW
U+3240 ㉀: uni3240-TW
U+3243 ㉃: uni3243-CN
U+3245 ㉅: uni3245-CN
U+3246 ㉆: uni3246-CN
U+3247 ㉇: uni3247-CN, uni3247-TW
U+3283 ㊃: uni3283-CN
U+3285 ㊅: uni3285-CN
U+3286 ㊆: uni3286-TW
U+3287 ㊇: uni3287-CN, uni3287-TW
U+328B ㊋: uni328B-TW
U+328E ㊎: uni328E-CN
U+3292 ㊒: uni3292-TW
U+3293 ㊓: uni3293-CN
U+3295 ㊕: uni3295-TW
U+3297 ㊗: uni3297-CN
U+329B ㊛: uni329B-CN, uni329B-TW
U+329C ㊜: uni329C-CN, uni329C-TW
U+329D ㊝: uni329D-TW
U+329E ㊞: uni329E-CN, uni329E-TW
U+32A2 ㊢: uni32A2-CN, uni32A2-TW
U+32A9 ㊩: uni32A9-CN, uni32A9-TW
U+32AA ㊪: uni32AA-CN, uni32AA-TW
U+32AC ㊬: uni32AC-CN
U+32AD ㊭: uni32AD-CN
U+32AE ㊮: uni32AE-TW
U+32B0 ㊰: uni32B0-CN, uni32B0-TW
I also found some other problems in these glyphs. For instance, HK version's ㈧ should map to uni3227-CN (currently mapped to uni3227-TW), and HK version's U+323E ㈾, U+3240 ㉀ should use the JP glyphs. Also, uni3247-TW has an incorrect "爭" component. But, if you ask me, instead of fixing these glitches, I'd rather suggest to remove the non-JP glyphs for all these characters except those with enclosed numerics, and release the CIDs for more meaningful uses. In short, I suggest to only keep these localized glyphs:
U+3223 ㈣: uni3223-CN
U+3225 ㈥: uni3225-CN
U+3226 ㈦: uni3226-TW
U+3227 ㈧: uni3227-CN, uni3227-TW
U+3283 ㊃: uni3283-CN
U+3285 ㊅: uni3285-CN
U+3286 ㊆: uni3286-TW
U+3287 ㊇: uni3287-CN, uni3287-TW
My rationale is that characters like ㈳, ㊭ are of Japan's legacy and pretty much not used by the Chinese communities. This is probably why the problems left unnoticed for so long (I happened to discover this when inspecting the outline disappearence issue caused by tx
). This also applies to the non-JP glyphs in U+1F200-U+1F2FF, and U+337F ㍿ (uni337F-CN and uni337F-CN-V) which is in the CJK Compatibility block.
(Edit: to be clear, I just mean I'm fine if these glyphs are nuked because the non-numeric enclosed glyphs don't much meaning to non-JP regions. But it's also fine keeping and fixing the localized glyphs. After all, these characters might already be in use by users outside JP because they are in Unicode, so there's nothing wrong providing localized forms)