My curated issues on Source Han Serif v2.001

[tamcy]

First of all, congratulations and thank you for the hard work on the release of Source Han Serif v2.001! Here I am listing the issues I found as of version 2.001, which I think are all isolated cases.

Here I try to retrain from writing too many remarks if I think the sample image is enough to illustrate the problem, so please tell me if they are not obvious enough.

[tamcy]

Section A. Remap suggestions

For the items below, I suggest to remap the affected regions to another glyph to better match the required regional standard (or the one chosen by the font).

(Click to view items already fixed in v2.003)

Issues affecting CN, TW and HK

Click here to for the image

#	Codepoint	Character	Affected Regions	Remap Suggestion
A1	U+5E3D	帽	CN,TW,HK	uni5E3DuE0101-JP (Fixed in v2.003)

Issues affecting TW and HK

Click here to for the image

#	Codepoint	Character	Affected Regions	Remap Suggestion	Additional remarks
A2	U+540D	名	TW,HK	uni540D-CN	The CN form which 夕 and 口 look visually centered looks more native to Chinese users
A3	U+62CB	拋	TW,HK	uni62CB-JP	I am aware of the 尢 design differences/indifferences between regions, but this is an isolated issue - the CN form which design of 尢 looks too close to 九 isn't very familiar to Traditional Chinese users (and we already have U+629B 抛) . So remapping the TW and HK regions to the JP glyph is more preferrable. (Fixed in v2.003)
A4	U+6F74	潴	TW^,HK	uni6F74uE0101-JP	(Fixed in v2.003)
A5	U+7239	爹	TW,HK	uni7239-CN	(Fixed in v2.002)

^: HKSCS characters, out of scope for TW. Remap suggested anyway for better regional consistency.

Issues affecting HK

Click here to for the image

#	Codepoint	Character	Affected Regions	Remap Suggestion	Additional remarks
A6	U+588F	墏	HK	uni588F-CN	(fixed in v2.003)
A7	U+5C07	將	HK	uni5C07-CN	(fixed in v2.003)
A8	U+5C4D	屍	HK	uni5C4D-CN	See the remarks in "死" (U+6B7B) below
A9	U+5D88	嶈	HK	uni5D88-CN	(fixed in v2.003)
A10	U+613F	愿	HK	uni613F-JP	(fixed in v2.003)
A11	U+63C3	揃	HK	uni63C3-JP90-JP	(fixed in v2.003)
A12	U+64E5	擥	HK	uni64E5-TW	(fixed in v2.003)
A14	U+69F3	槳	HK	uni69F3-CN	(fixed in v2.003)
A15	U+6B7B	死	HK	uni6B7B-CN	The CN form which the 丿 stroke in 匕 is placed lower looks more native to Traditional Chinese users. Also it's common to see "歹" and "匕" connected together. Other HK dedicated glyphs with this component also adopts the connected form (examples).
A16	U+6F3F	漿	HK	uni6F3F-CN	(fixed in v2.003)
A17	U+714E	煎	HK	uni714E-JP90-JP	(fixed in v2.003)
A18	U+731C	猜	HK	uni731CuE0101-JP	(fixed in v2.003)
A19	U+76E5	盥	HK	uni76E5-JP	(fixed in v2.003)
A20	U+787B	硻	HK	uni787B-CN	(fixed in v2.003)
A21	U+86F8	蛸	HK	uni86F8-JP90-JP	(fixed in v2.003)
A22	U+87BF	螿	HK	uni87BF-CN	(fixed in v2.003)
A23	U+8CA7	貧	HK	uni8CA7-CN	Minor difference at the second stroke of 八 - in this font, HK follows CN's form
A24	U+8E61	蹡	HK	uni8E61-CN	(fixed in v2.003)
A25	U+9306	錆	HK	uni9306-JP90-JP	(fixed in v2.003)
A26	U+962A	阪	HK	uni962AuE0101-JP	(fixed in v2.003)
A27	U+975C	靜	HK	uni975CuE0102-JP	(fixed in v2.003)
A28	U+9BB9	鮹	HK	uni9BB9uE0101-JP	(fixed in v2.003)
A29	U+9BD6	鯖	HK	uni9BD6-JP90-JP	(fixed in v2.003)

Issues affecting HK

Click here to for the image

(The items listed below are mostly fixed. The only remaining item is A13, currently the HK form is mapped to the TW CID)

#	Codepoint	Character	Affected Regions	Remap Suggestion	Additional remarks
A13	U+68C8	棈	HK	uni68C8uE0101-JP

[tamcy]

Section B. New Glyph Assignment Suggestions

For the following items, I believe a new glyph would be needed for the glyph to match the required regional standard or the one chosen by the font.

Items affecting CN, TW and HK

Click here to see the image

#	Codepoint	Character	Affected Regions	Current Glyph Name	Reason
B1	U+99F5	駵	CN,TW^,HK	uni99F5-JP	Incompatible stroke form for "匚" in 丣, it should be the same as the stroke form of "匚" in "亞".
B2	U+6F7B	潻	CN,TW,HK	uni6F7B-JP	Last stroke of 禾 is different

Items affecting TW and HK

Click here to see the image

#	Codepoint	Character	Affected Regions	Current Glyph Name	Reason
B3	U+5DAF	嶯	TW,HK	uni5DAF-CN	Different design for 耳
B4	U+69B3	榳	TW,HK	uni69B3-CN	Different design for 廷 *
B5	U+7407	琇	TW,HK	uni7407-CN	Last stroke of 禾 is different
B6	U+78F0	磰	TW^,HK	uni78F0-JP	Vertical stroke of 𦍌 doesn't protrude
B7	U+7AE7	竧	TW,HK	uni7AE7-CN	Incompatible design for 身
B8	U+8A94	誔	TW^,HK	uni8A94-CN	Incompatible design of 廷 *
B9	U+8B4F	譏	TW,HK	uni8B4F-CN	The 人 component doens't protrude the horizontal stroke (all other 幾-composed characters have their dedicated TW/HK glyphs (examples); U+8B4F 譏 is the only exception)
B10	U+8E8D	躍	TW,HK	uni8E8D-CN	Hooks exist for 羽
B11	U+9385	鎅	TW^,HK	uni9385-CN	Different design for 界 (Other similar characters for reference)

• A note about B4 (U+69B3 榳) and B8 (U+8A94 誔): Interestingly the CN glyph in the Unicode code chart also shows the lower part of 𡈼 as 土, not 士. But I don't know much about the GB standard, so I still treat them differently.

Items affecting HK

Click here to see the image

#	Codepoint	Character	Affected Regions	Current Glyph Name	Reason
B12	U+3C18	㰘	HK	uni3C18-CN	Different design for 糸
B13	U+5F5D	彝	HK	uni5F5D-CN	Different design for 糸
B14	U+655C	敜	HK	uni655C-TW	Second stroke of 人 is different
B15	U+6F78	潸	HK	uni69B3-CN	Last stroke of the second 木 is different
B16	U+7298	犘	HK	uni7298-CN	Last stroke of the second 木 is different
B17	U+8EB2	躲	HK	uni8EB2-CN	Different design for 身

^: HKSCS characters, out of scope for TW. Remap suggested anyway for better regional consistency.

[tamcy]

Section C. Glyph Design Update Suggestions

This is similar to section B, but fixing the item doesn't involve a new glyph be added. I will also list the concept fix if the illustrated problem may not be obvious.

Click here to for the image

#	Codepoint	Character	Affected Glyph Name	Remarks	Concept (if available)
C1	U+5C7E	屾	uni5C7E-TW	The composition should look like 昍, i.e. two components standing beside each other.
C2	U+515C	兜	uni515C-HK	The second stroke of the topleft component ("└") is written as one stroke.
C3	U+615D	慝	uni615D-TW	Adjust the 艹 component to match the v2.000 form (vertical strokes are not slanted)	(Fixed in v2.003)
C4	U+7661	癡	uni7661-TW	The 匕 component is not correct.	(Fixed in v2.003)
C5	U+88CA	裊	uni88CA-TW	The left vertical stroke above 衣 should not be connected with the 𠃌 stroke. For this component the same form for JP or CN could be used.	(Fixed in v2.003)
C6	U+23256	𣉖	u23256-HK	The 火 component is different from the chosen HK form.

That's it, thanks again!

[Marcus98T]

My additional notes:

A3 might probably need to have the JP glyph on all regions as per this issue. Then the CN glyph can be removed.
A4 will need uni6F74uE0101-JP to map to CN as well, because like HK and TW, there is no extra dot in 者 for the CN region.

A26 appears to have an EDB reference error showing 阱 instead of 阪. Maybe @tamcy needs to edit the graphic and reupload.

B11 will also need the new glyph to be mapped to JP and KR for consistency sake.

EDIT: Funny that the CN region for 界 components is mapped to JP for both Sans and Serif. I suggest instead to adjust those two characters: 䛺 (U+46FA) and 鎅 (U+9385) to be more consistent with the other 界 characters. Don't add any new glyphs and don't change the mappings for B11.

As for 琾 (U+743E), restore the v1 JP glyph to replace the CN glyph.

For C1, the TW form, after adjustment, would also need to be mapped to JP and KR for aesthetics sake.

Now onto the topic of unifying glyphs for the sake of reducing unnecessary CIDs (as far as I can without breaking China's 新字形 rule):

As for A2, in the future I might need to make a new issue on both Sans and Serif issue pages concerning a potential unification of the 夕 and 多 components, on whether it's acceptable if the dot (丶) can touch the ク part for the CN region. 名 (U+540D) however, could be handled differently because of balance, so for now, I agree for A2 to be mapped to CN for TW/HK regions.

As for A8 and A15, same as A2 in terms of making a new issue, except it has to do with the 死 component, where all regions may need to have 匕 connect to 一 (Of course I know the TW region have to be treated differently for the 匕 part, so not entirely unifiable, and yes, unlike the JP style that I prefer for other components, the CN style would be preferred for JP and KR regions).

A23, for the time being, I agree that the HK region should map to the CN form, but on whether the 八 component can be unified with CN dropping that Song-style detail from the second stroke in favour of the JP-style glyph shape, I wonder if it could be possible without breaking the 新字形 rule? That I'll post a new issue soon.

UPDATE x2: I have made a new consolidated issue page concerning glyph unification.

[tamcy]

(All items in this comment have been fixed in v2.003. Click to view.)

An update on 2022/3/3:

As per HK's reference glyph and the stroke sequence shown in the Education Bureau website, the middle vertical stroke of 畢 is rendered/written as a single stroke, not two.

Thus, for HK and TW, the standard/recommended stroke order/stroke form of this component differs in two ways:

1. In HK, 日 is written first, followed by two 十 (four strokes), then a 二 (top stroke is longer), and finally a 丨 that strikes through the 日 component.
2. in TW, 田 is written first, followed by the lower parts, which are 卄 (three strokes), and then a 二 (top stroke is longer), and finally the 丨 which starts below the 田.

However, 3 (out of 17) HK glyphs are found to have the middle vertical stroke separated:

So I would like to suggest to have these 3 glyphs corrected:

#	Codepoint	Character	Affected Glyph Name
C7	U+7562	畢	uni7562-HK (fixed in v2.003)
C8	U+7BF3	篳	uni7BF3-HK (fixed in v2.003)
C9	U+7F7C	罼	uni7F7C-HK (fixed in v2.003)

Still on the same component, I found a possible regression of U+55F6 嗶 and U+6EED 滭 for TW:

In v2.000 the 2 codepoints shared the JP glyph in v2.000, which is fine. In v2.001 they have their own glyph instances, but the 畢 component is now incorrect - the 田 now becomes a 日.

#	Codepoint	Character	Affected Glyph Name
C10	U+55F6	嗶	uni55F6-TW (fixed in v2.003)
C11	U+6EED	滭	uni6EED-TW (fixed in v2.003)

Thanks!

[tamcy]

U+7CEF: Currently mapped to CN, but the component design of 雨 should be the same as that of TW. The TW glyph cannot be used either because of the design difference for 米.

#	Codepoint	Character	Affected Regions	Current Glyph Name	Reason
B13	U+7CEF	糯	HK	uni7CEF-CN	Different design for 雨

[tamcy]

(also applies to v2.003)

U+7E6D: The TW glyph is used for HK but the design of 糸 is different.

#	Codepoint	Character	Affected Regions	Current Glyph Name	Reason
B14	U+7E6D	繭	HK	uni7E6D-TW	Different design for 糸