source-han-serif
source-han-serif copied to clipboard
adobe-fonts
Redesign suggestion for the middle "=|="-like component in 屬/褱/犀 etc.
Source Han Serif version: v2.000 Affected characters: Those with either one of the following components: 屬/脊/犀/皐/臯/褱. Affected regions: CN, TW, HK
Summary
I raised a similar suggestion in the Source Han Sans issue tracker three years ago: https://github.com/adobe-fonts/source-han-sans/issues/204#issuecomment-446537524
The suggestion is to modify the
and
components, so that the four dot-like strokes do not necessary touch the other strokes. Like this:
Details
For the "=|=" and "=人="-like components in this issue, I'll call D1-D4 strokes the "dot strokes". I know D2 and D3 aren't really dots, this is just for simplicity.
I am not sure why the original designer seems so insistent on having D3 and D4 to touch each other, and having D2 and D3 to touch the center component.
For TW and HK, having them separated is the norm, and I think no further explanation is needed.
So, given the consistency and insistency, I first thought it might be due to the standard set by GB18030. But turns out it isn't. I've checked the PRC government's stroke order document 通用規範漢字筆順規範 released on March 2021, and confirmed that these four dots are counted as four separate strokes, so having them sticked together shouldn't be neccessary.
Here is the excerpt of 犀 from the document (P.353). Notice how the four "dots" are rendered, and that they are counted as four distinct strokes.
Another example is the character 脊 on page 220, the four dots are disconnected from 人:
While this document is published recently, I also checked several commerical Simplfied Chinese fonts and some Simplified Chinese publications released before 2021. Still I could spot the "untouched strokes" design.
This is SimSun, the Simplified Chinese font bundled with Windows 10:
And here are the samples from commercial fonts:
To conclude, I believe that CN glyphs are not required to be designed in a way that the dots look inseparable.
The problem
The problem is that the current design seems to bring more harm than good. Just like at the following three characters:
Some characters, like 穉 (U+7A49), tries to compensate this by moving strokes D3 and D4 up. This still looks subpar, because of the empty space difference below D2 and D3:
It can also lead to weirdlooking design:
I know, "weird looking" can sound subjective. But a few HK/TW characters in Source Han Serif are designed using the "separated form". Which is good for comparison:
I am not saying that the four dots should never touch each other or the vertical stroke (this is somewhat inevitable in the heaviest weight). The point is that, having D2 and D4 to touch the middle component and having D4 touched by D3 shoud not be the highest command, and the design of the component should not be bound by this requirement.
One thing worth noting is that characters with the component 屬 and 褱 are not used by Simplified Chinese users daily - they use 属 in place of 屬. 褱 is deemed uncommon, and is simplified to 不 when used as a component (懷→怀, 壞→坏). Which is why you won't find examples of these two components in the document.
Affected glyphs
Here is the list of to-be-designed glyphs if this proposal is accepted. Struck-out glyph names are those that already conform to the proposed form and need no adjustment.
| # | Codepoint | Character | CN | TW | HK |
|---|---|---|---|---|---|
| 1 | U+34FE | 㓾 | uni34FE-CN | ||
| 2 | U+3728 | 㜨 | uni3728-CN | ||
| 3 | U+3733 | 㜳 | uni3733-CN | uni3733-HK | |
| 4 | U+40F6 | 䃶 | uni40F6-CN | ||
| 5 | U+4188 | 䆈 | uni4188-CN | ||
| 6 | U+4335 | 䌵 | uni4335-CN | ||
| 7 | U+4659 | 䙙 | uni4659-CN | ||
| 8 | U+4671 | 䙱 | uni4671-CN | ||
| 9 | U+4704 | 䜄 | uni4704-CN | ||
| 10 | U+4765 | 䝥 | uni4765-CN | ||
| 11 | U+4831 | 䠱 | uni4831-CN | ||
| 12 | U+4C2A | 䰪 | uni4C2A-CN | ||
| 13 | U+529A | 劚 | uni529A-CN | ||
| 14 | U+5637 | 嘷 | uni5637-CN | ||
| 15 | U+5651 | 噑 | uni5651-CN | ||
| 16 | U+56D1 | 囑 | uni56D1-CN | ||
| 17 | U+5849 | 塉 | uni5849-CN | ~~uni5849-TW~~ | |
| 18 | U+5880 | 墀 | uni5880-CN | ||
| 19 | U+58DE | 壞 | uni58DE-CN | ||
| 20 | U+5B4E | 孎 | uni5B4E-CN | uni5B4E-TW | |
| 21 | U+5C6C | 屬 | uni5C6C-CN | ||
| 22 | U+5D74 | 嵴 | uni5D74-CN | ~~uni5D74-HK~~ | |
| 23 | U+5FB2 | 徲 | uni5FB2-CN | ||
| 24 | U+61F7 | 懷 | uni61F7-CN | ||
| 25 | U+6468 | 摨 | uni6468-CN | ||
| 26 | U+65B8 | 斸 | uni65B8-CN | ||
| 27 | U+66AD | 暭 | uni66AD-CN | ||
| 28 | U+66CD | 曍 | uni66CD-CN | ||
| 29 | U+66EF | 曯 | uni66EF-CN | ||
| 30 | U+69F9 | 槹 | uni69F9-CN | ~~uni69F9-HK~~ | |
| 31 | U+6A28 | 樨 | uni6A28-CN | ||
| 32 | U+6A70 | 橰 | uni6A70-CN | ||
| 33 | U+6AF0 | 櫰 | uni6AF0-CN | ||
| 34 | U+6B18 | 欘 | uni6B18-CN | ||
| 35 | U+6F3D | 漽 | uni6F3D-CN | ||
| 36 | U+7024 | 瀤 | uni7024-CN | ||
| 37 | U+705F | 灟 | uni705F-CN | ||
| 38 | U+7225 | 爥 | uni7225-CN | uni7225-HK | |
| 39 | U+7280 | 犀 | uni7280-CN | ||
| 40 | U+734B | 獋 | uni734B-CN | ||
| 41 | U+7354 | 獔 | uni7354-CN | ||
| 42 | U+74CC | 瓌 | uni74CC-CN | ||
| 43 | U+7620 | 瘠 | uni7620-CN | ~~uni7620-TW~~ | |
| 44 | U+7690 | 皐 | uni7690-CN | ~~uni7690-HK~~ | |
| 45 | U+76A1 | 皡 | uni76A1-CN | ||
| 46 | U+76A5 | 皥 | uni76A5-CN | ||
| 47 | U+77DA | 矚 | uni77DA-CN | ||
| 48 | U+7A49 | 穉 | uni7A49-CN | ||
| 49 | U+7FF6 | 翶 | uni7FF6-CN | ||
| 50 | U+7FFA | 翺 | uni7FFA-CN | ~~uni7FFA-HK~~ | |
| 51 | U+8032 | 耲 | uni8032-CN | ||
| 52 | U+810A | 脊 | uni810A-CN | ~~uni810A-TW~~ | |
| 53 | U+818C | 膌 | uni818C-CN | ~~uni818C-TW~~ | ~~uni818C-HK~~ |
| 54 | U+81EF | 臯 | uni81EF-CN | ||
| 55 | U+8639 | 蘹 | uni8639-CN | ~~uni8639-TW~~ | |
| 56 | U+863E | 蘾 | uni863E-CN | ~~uni863E-TW~~ | |
| 57 | U+883E | 蠾 | uni883E-CN | ||
| 58 | U+8931 | 褱 | uni8931-CN | ||
| 59 | U+8E50 | 蹐 | uni8E50-CN | ~~uni8E50-TW~~ | |
| 60 | U+9072 | 遲 | uni9072-CN | uni9072-TW | |
| 61 | U+9483 | 钃 | uni9483-CN | ||
| 62 | U+9DBA | 鶺 | uni9DBA-CN | ~~uni9DBA-TW~~ | |
| 63 | U+9DF1 | 鷱 | uni9DF1-CN | ||
| 64 | U+9E61 | 鹡 | uni9E61-CN | ||
| 65 | U+24161 | 𤅡 | ~~u24161-HK~~ | ||
| 66 | U+28B2F | 𨬯 | ~~u28B2F-HK~~ |
Affected glyphs: 66 (CN) + 2 (TW) + 2 (HK) = 70.
Another option would be to leave the current CN form as-is but have a separate version for TW/HK glyphs. This requires 34 new CIDs, and 3 existing TW/HK glyphs be tweaked.
Conclusion
This is not an aesthetics versus standardized form issue. In fact, the proposed designed conforms better to the reference glyphs published by TW and HK, and I believe this is also true for CN. Given the glyphs are shared among CN,TW and HK, I believe this is a change that will be benefitial to users and the overall quality of the font.
Thank you.
I also want to note 泰/康/彔/录/隶 is a different issue altogether and the handling by SHS is not consistent within the same region either.
This is something that has been bothering me, too, so thank you for writing it up so nicely. I will discuss with the designers and see what everyone thinks.
