Problem with Sans JP: 陋 U+964B is incorrect

[NightFurySL2001]

Reported in #556 (comment) but separated to prevent tracking TC and JP together in an issue.

陋 U+964B at Adobe-Japan1 CID+7094 is incorrect for JP presumeable copied from Source Han Sans which has the same problem.

Adobe-Japan1:

[SCLu17]

This seems to be a non-issue where Japan allows for design variances in terms of stroke directionality (Agency for Cultural Affairs, Jōyō Kanji Hyō, “Regarding Mincho Designs”, p. 3 and relevant example in “3 (5)”, p. 5). Comparing different Gothics, we can see the last stroke of the 丙 component isn’t standardized across different typefaces (glyphs which use the 丶 form are highlighted with a black dot underneath):

The Plex designers seem to be using Kozuka Gothic/Source Han Sans as a reference.
Even if the designers are following Adobe-Japan1 forms (which uses Kozuka Mincho for its representative glyphs), Adobe’s own documentation states that the glyphs are meant as references, not standards, i.e. there is no need to strictly adhere to the above forms, which brings me back to the Japanese government’s comment on typeface design variations.
As this is a permissible design variation, it would be excessive to label the design as “incorrect”. Consistency in the 丙 component (one form or the other) is optional and shouldn’t be prioritized.

[mjabbink]

The response from Sandoll designers

For this Issue, we can select all of the following methods:
1.          Maintain the current design.
2.          Modify only 人 of the 陋(U+964B) to '丿+丶'.
3.          Unify all the 人 of the  丙  to '丿 + 丶'.
 
It would be better to modify only for 陋(U+964B).
This is my opinion, and you can judge which would be better. I will change the design to what you have confirmed among the three.
 
Below is an additional description:
Sandoll team followed the Kozuka font series a lot when we set the shape criteria for IBM Plex JP.
The Kozuka series is an early digital font family widely believed to have faithfully followed the font Heisei Mincho used as a reference in the Japanese Industrial Standard (JIS).
Part of these is now shared as open source, and many Japanese fonts that are considered legitimate also frequently refer to this typeface.
 
However, the Kozuka series chose a different design for the '陋' than Heisei Mincho.
Unlike GB(Chinese standard), JIS is not obligated to have a specific design [Attachment 1, p.4-5]. However, the current design is indeed found in a small number of various Japanese documents and fonts.
In conclusion, the current design is not wrong. Still, it would be better to modify this font to preserve its shape authenticity and make it more clearly IBM Plex-like.
 
On the other hand, 昞炳 is currently recommended for symmetrical design because it is registered in JIS as a design corresponding to "旧字体(Old-type)" among the two types (Old-type or New-type) that distinguish Japanese Kanji from the times. Of course, it is not an obligation, so the direction in which all 丙 designs are unified is not wrong.