-
Notifications
You must be signed in to change notification settings - Fork 1.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Standardize TW/HK 八 and 入 components with JP/KR forms #414
Comments
Here are the tables for the 全 component to consolidate the glyphs on. Sans![]() LEGEND
Serif![]() LEGEND
|
I won't do a table for the 㒼 component because there are only a few characters that have issues. Both Sans and Serif![]() ![]() Remap 鬗 (U+9B17) and 襔 (U+8954) to the TW glyph for JP and KR (marked in magenta). A side note is that there was a v1 JP glyph for 鬗 (U+9B17) in Serif, but it's off-scope so it cannot be restored. Sans![]() Adjust the TW and HK glyphs for 濷 (U+6FF7) so as to match the JP shape, with regards to the 入 parts (in yellow, circled), so as to make it consistent with the other 㒼 components (although admittedly they are inconsistent within themselves, even in the JP glyphs, see 鏋 above). ![]() Then you can remap to the HK glyph for the JP/KR locales (marked in magenta). Alternatively, Adobe can use an unreleased JP glyph that never made it to v1 and remove the HK glyph, and then it can be used for the JP/KR/HK locales. If that happens, the TW glyph must also be adjusted for consistency. Serif![]() Restore v1 JP (in green) to replace the HK glyph for 濷 (U+6FF7, in red), then it can be safely mapped to JP/KR/HK (in blue). If Adobe can find unreleased/removed JP glyphs that suit the TW and/or HK locales, they could have used them rather than get Arphic to redesign a similar-looking glyph, it's a waste of effort. |
Here are the tables for the 八 component to consolidate the glyphs on (does not apply to the HK locale). For now, the characters that contain 公 or 分 components are not included. Sans![]() LEGEND
Serif![]() LEGEND
|
Here are the tables for the 公 component to consolidate the glyphs on (does not apply to the HK locale). Because the bottom part of 公 is not unifiable with JP/KR forms, all the TW glyphs would just need adjusting on the top part to match JP/KR forms. The tables apply to both Sans and Serif. ![]() ![]() Adjust TW glyph to match JP/KR forms (marked in yellow)
Remap to TW glyph for JP/KR (marked in magenta)It will be better for some out-of-scope glyphs to at least follow a bit more of the faux-traditional orthography so I like to remap the following glyphs from CN to TW for JP/KR.
|
Here are the tables for the 穴 component to consolidate the glyphs on (does not apply to the HK locale). ![]() Both Sans and SerifAdjust TW glyph to match JP/KR forms (marked in yellow)
Remap to TW glyph for JP/KR (marked in magenta)
Also of note, there is a JP glyph for 袕 (U+8895) in Serif (marked in dark purple) which is off-scope, but it is not removed as of v2. Either keep it, or remove it and remap to the TW glyph. UPDATE 2023-09-08: Unfortunately I missed 屄 (U+5C44). I have updated the tables and graphic to include that character and will close the other issue for consolidation purposes. |
Here are the tables for the 分 component to consolidate the glyphs on (does not apply to the HK locale). Sans![]() LEGEND
Serif![]() Red - remove TW glyph and use JP/KR glyph instead
I think that's the last of the 八 and 入 components. EDIT 2023-08-11: Updated status of 掰 (U+63B0) and 椕 (U+6915). |
I propose making the TW/HK forms of 八 (TW only) and 入 (especially the length of the top horizontal stroke that is connected to the throw stroke on the right) follow that of the JP/KR designs. This will make the glyphs more consistent and ensures better readability.
This will involve adjusting some TW glyphs, and even restoring some v1 JP glyphs (or adding JP glyphs that was designed but never made it to v1) to replace the current TW glyph if needed.
Doing this will also free up some CIDs as well, probably in the hundreds.
Here are the examples of the discrepancy between the length of the horizontal strokes between JP/KR and TW/HK, in Sans and Serif.


For the 公 component, all the TW glyphs containing this should be manually adjusted to match the balance of JP/KR glyphs because the bottom 厶 part is not unifiable.
Again, I’m giving a commercial font as a reference for why I am proposing this change, even though by right we should not copy another typeface. Hiragino Sans CNS (TW) keeps the length of the 八 and 入 components similar to the JP locale (if set to JIS78 forms).

Also, quoting from this issue on Serif:
I will list the glyphs affected in the coming weeks, along with a separate version for Serif. I did list glyphs for the 全 component in the same issue page as above, but I will create a more comprehensive list soon which will also include 八.
The text was updated successfully, but these errors were encountered: