Unify to 𡪽 (U+21ABD)?

Both characters are variants of "窺". Based on UCV rule #354i, TCA agrees to unify to 𡪽 (U+21ABD).

On the issue of unification to 𡪽 U+21ABD:

The following 5 characters are encoded containing the 䂓 component:

𡪽 U+21ABD - Extension B, "canonical" form being suggested for separate encoding
𢵅 U+22D45 - Extension B, "canonical" form encoded as 摫 U+646B
𥨖 U+25A16 - Extension B, "canonical" form encoded as 窺 U+7ABA
𧡯 U+2786F - Extension B, error form of 瞡, canonical form "⿰月䂓" not encoded
𭬉 U+2DB09 - Extension F, "canonical" form encoded as 槻 U+69FB

𡪽 U+21ABD itself is an error form of 窺 U+7ABA.

If we apply the principle that "a 'canonical' form must be encoded separately from an existing variant form", then the pseudo-canonical form for U+2786F would also be encoded separately.

This argument of "a 'canonical' form must be encoded separately from an existing variant form as a glyph change will destabilize existing data sources" has been pushed by China multiple times and it has been applied in some previous working sets. At least one character has been disunified because the variant form was found in the GKX, as GKX was considered authoritative.

Per my memory this is the first time it is being applied to a character from GHZ origin. Historically China has opted to modify some character glyphs for those with the GHZ origin, especially the canonical glyph was found in GHZR.

As GHZ is also considered an authoritative dictionary by IRG, same argument could be made. If IRG accepts the idea of "a 'canonical' form must be encoded separately from an existing variant form", we should limit it to only characters where the existing encoded character is encoded from specific authoritative sources.

What is regarded as a canonical form may also change based on who is looking at it. So if this principle is accepted, every time this principle is invoked, IRG should keep a record of the selected form, such that future characters are normalized to prevent duplicated forms from being encoded.

--

On the issue of the general unifiability of 規 and 䂓:

This was added from the discussion of 00543.

A number of variant forms have already been registered on zi.tools:


(first character and third character)



Other examples from the MOE Variant Dictionary:



I prefer to keep this UCV.

［元］王惲 著 ，楊亮、鍾彦飛 點校 ：《王惲全集彙校》，《中國古典文學基本叢書》，中華書局，2013年11月，第1版，第121頁。

［宋］李昉 等 撰 ：《太平御覽》，中華書局，1960年02月，第1版（→商務印書館1935影宋本），第4086頁。

［宋］李昉 等 編 ：《文苑英華》，中華書局，1966年05月，第1版，第220頁。

［宋］李昉 等 編 ：《文苑英華》，中華書局，1966年05月，第1版，第1064頁。

［清］瞿鏞 撰 ：《鐵琴銅劍樓藏書目録》，《宋元明清書目題跋叢刊》，中華書局，2006年06月，第1版，第106頁。

［清］董誥 等 編 ：《全唐文》，中華書局，1983年11月，第1版，第10947頁。

《干禄字書一卷》,明萬曆中刊本

［清］羅振玉 箋, 《干禄字書箋證一卷》, 貞松老人遺稿本

《干祿字書一卷》, 《後知不足齋叢書》清光緒中常熟鮑氏刊本

［清］繆荃孫 著 ，張廷銀、朱玉麒 主編 ：《繆荃孫全集 金石》，鳳凰出版社，2014年08月，第1版，第90頁。

戴應新 編著 ：《長安鳳棲原韋氏家族墓地墓志輯考》，三秦出版社，2021年03月，第1版，第199頁。