Single example of an idiosyncratic way of writing 窓, not sufficient evidence for encoding. Unify with 𥦗 by UCV #22.

The radical of 𥦗 is 穴 while the radical of ⿱宀总 is 宀, which means the change is important enough.

《中华姓氏源流大辞典》(中华书局，2014年) p. 1360

Based on the new evidence, I still believe that unification with U+25997 𥦗 is appropriate.

Agree on unification to 𥦗 (U+25997).

Support the unification to 𥦗 (U+25997).

It’s common to write 穴 (⿱宀八/⿱宀儿) as ⿱宀丷 in the hand writing.

Please provide some evidences to prove " it is common to write 穴 (⿱宀八/⿱宀儿) as ⿱宀丷 in the hand writing".

I understand Andrew's opinion about this unification case. What I want to say here is, the unification rule could applicable to most cases, except for special circumstance. Changing 八/儿 to 丷 do also change the cognitive style about the whole character. People always look upon a thing with the simplest thinking. A most ordinary Chinese can only think that GDM-00313 is combined by 宀 and 总, unless he/she has a high degree of expertise. Rules should not go against common sense.

綺樓重夢四十八回
文津阁四库全书·乐府指迷
愛日吟廬書畫錄
東坡集

新訂中州全韻

This shape is seen in many books, used as last name and used in place names as well.
徐铁生: 《中华姓氏源流大辞典》,中华书局出版发行,北京市白帆印务有限公司，2014年1月北京第1版，2014年1月北京第1次印刷,ISBN978-7-101-09024-6, page1360

SJT 11239-2001 信息技术 信息交换用汉字编码字符集 第八辅助集

Considering the unification relates to the change of radical, I think it is better to encode the character seperately.

Still unify to 𥦗 (U+25997) by UCV #22, because no actual use of abstract shape ⿱[宀][总] is found.
If the actual shape ⿱|宀||总| is regarded more important, it's okay for G-source to update the reference glyph of 𥦗 (U+25997).

Also, as a premature point of view, the so-called common sense is usually not an a priori self-unfolding of intellectual intuition, but remains a product mediated by rules. Thus, rather than confronting the common sense, the rule is a statement of what the common sense really is, so that when the common sense is used to make analytic judgments, there is already a set of unspoken and well-spoken rules at work behind the scenes. It's just that this set of rules is not fully compatible with the current rules.

BTW, the character is deleted from Extention I because that it is in IRG WS2021. It is not proper to unify the character after China deciding to remove it from Extention I.
What's more, I‘d like to say that we can not live in theory and the ISO 10646 standard should be made for practical usage. The fact that the character is usually treated as different character from 𥦗 (U+25997) reflects the wide demand of using all two characters in the same context. Unifying this kind of common variants will cause that people use PUA to represent it, which can be easily avoided if we encode them seperately.

Considering:
1) the radical of ⿱宀总 is 宀 while the radical of 𥦗 (U+25997) is 穴;
2) the variant ⿱宀总 is used in many books and 𥦗 (U+25997) as well;
3) ⿱宀总 is used as last name, which is important to people;
4) ⿱宀总 is used in Chinese place names;
5) ⿱宀总 is deleted from Extention I by China because it is in IRG WS2021;
6) people will normally treate ⿱宀总 and 𥦗 (U+25997) as different characters.
It is obviously better to encode ⿱宀总 seperately rather than unify it based on academic theory or other rigid “rules”.

This form was used commonly in 四庫全書.


▲ 《朱子語類》, 文淵閣四庫全書本, 卷五十三, folio 12A


▲ 《叶韻彙輯》, 文淵閣四庫全書本, 卷十二, folio 12A


▲ 《叶韻彙輯》, 文淵閣四庫全書本, 卷二, folio 1A

Maybe we could add 穴 and ⿱宀丷 as a new UCV.

空 is written as ⿳宀丷工, but we also identify the radical as 穴 not 宀, that means the radical for this character is still 穴 not 宀.

As the experts’ comments, two forms are both needed in China, so maybe the best way is to use IVS not PUA for all the end-users.

▲ 北京书同文数字化技术有限公司, 古籍汉字字频统计. 商务印书馆, 2008: page 120

The G4K database captures the character and does not submit it to Ext B. Consider that filtering has been done when generating the G4K repertoire.

We can't fantasize about a practice system that is advanced enough to support the latest repertoire and backward enough to not support the 20-year-old IVD system, such fantasies are only a fragile defense of some poor self-esteems.

The following four sources had not been posted as evidence. If no experts provided these, ORT manager should remove them.

宋元以来俗字谱
檮杌閑評
花嶼詞
樂府詩鈔

I strongly oppose to the comments in comment #14288 on behalf of The Center for Toponym Research of Sichuan In-ternational Studies University, especially the words "such fantasies are only a fragile defense of some poor self-esteems". This kind of sarcasm to national bodies should not be allowed in IRG.
To clarify, I am suggesting encoding the character seperately not just because some information processing systems don't support IVD. I am saying that, for this particular case, considering the facts I brought out in comment #14280, it is much better to encode ⿱宀总 seperately. For me, I will definately take part the character as ⿱宀总 at the first sight. What's more, the case of ⿳宀丷工 doesn't stand because ⿱丷工 is not a common character but ⿱丷𢗀 is.

The conclusion on IRG #58 is “Pending for more evidence”, since new evidences has been provided, there is no need to raise any new issues.
Moreover, this character is used as frequently as 900 times in my database, quite stable to be encoded.

Let me sort the discussion out for experts:
Experts diagree to encode ⿱宀总 seperately says:
1) Based on UCV#22, 八 and 丷 can be unified;

2)⿱宀总 should be take apart as ⿳宀丷𢗀 academically so ⿱宀丷 should be unified to 穴, and the radical of ⿱宀总 should be 穴(comment #14286).
Experts agree to encode ⿱宀总 seperately says:
1)Technically, ⿱宀总 and 𥦗 (U+25997) can be unified to each other;
2)Practically, most people would like to take part ⿱宀总 as ⿱宀总 because ⿱丷𢗀 is a very common character. Meanwhile, most people would like to take part 𥦗 (U+25997) as ⿳穴口心(i.e. ⿱穴𢗀). Then considering 宀 is not unifiable to 穴, 总 is not unifiable to 𢗀, so ⿱宀总 and 𥦗 (U+25997) should not be unified.
3)We think it is better to take part the character in a more practical way but take part the character academically thus they should not be unified.
4)Considering more:
a)The variant ⿱宀总 is used in many books and 𥦗 (U+25997) as well;
b) ⿱宀总 is used as last name, which is important to people;
c) ⿱宀总 is used in Chinese place names;
d) ⿱宀总 is deleted from Extention I by China because it is in IRG WS2021;
It is obviously better to encode ⿱宀总 seperately rather than unify it based on academic theory or other rigid “rules”.
Last but not least, I noticed that Eiso's comment in #14291 suggesting ORT manager should remove following source references:
宋元以来俗字谱
檮杌閑評
花嶼詞
樂府詩鈔
I'd like to say that we found totally more than 120 evidences for ⿱宀总, which are enough to prove point 4.a in this comment. I will post some of them in my next comment.

⿱宀总 is actually not a historic or very rare character:
Evidence1: 朱錦琮 撰：治經堂詩集，清道光四年（1824）刻本，卷四，page20

Evidence2: 田汝耔 撰：漢隸分韻，清乾隆三十八年（1773）九沙萬氏刻本，序言，page1

Evidence3: 錢大昭 撰：廣雅疏義，日本昭和十五年（1940）靜嘉堂影印清抄本，卷十三，page11

Evidence4: 李調元 輯：蜀雅，清乾隆中綿州李氏萬卷樓刻、嘉靖十四年（1809）李鼎元重校印本，卷二十，page16

Evidence5: 阮文藻 撰：聽松濤館詩鈔，清道光十一年(1831）刻本，卷六，page20

Evidence6: 御定佩文韻府，清乾隆間寫摛藻堂四庫全書薈要本，卷八十一，page42

Evidence7: 王桂 撰：葵書，清光緒六年(1880)刻本，卷上，page52

Evidence8: 新刻今古傳奇，清嘉慶二十三年(1818)刻本，卷十二，page6

Evidence9: 彭元端 撰：五代史記注，清道光八年(1828)刻本，卷五十四，page4

Evidence10: 劉一明 撰：西遊原旨，清嘉慶二十四年(1819)湖南常德同善分社刻本，卷五，page5

Evidence11: 沈濤 撰：常山貞石記，清光緒二十年(1894)靈溪精舍刻本，卷十三，page26

Evidence12: 張居正 撰：明張文忠公詩文集，清宣統三年(1911）醉古堂石印本，卷十一，page15

Evidence13: 王三接 撰 ，(明)王用言 輯：王槐溪先生文集五卷，明萬曆三十六年(1608)王學曾刻本，卷一，page14

Evidence14: 吴嵩梁 撰：香蘇山館全集，清道光二十三年(1843)刻本，卷四，page12

Evidence15: 祁东县志，中国文史出版社，1992年10月，page81


Also seen in the PUA font of Zhong Hua Book Company(中华书局):

1. What evidence is there for 𥦗? Could 𥦗 already be a normalization of ⿱宀总 that clarifies the relationship to 窓 by keeping the familiar shape of 穴 instead of introducing ⿱宀丷?
2. Why is 穴部, when it is composed of 宀(semantic)忩(phonetic), which is essentially equivalent to 宀(semantic)总(phonetic)?

What I mean on Comment #14291 is only for the editorial issue. If this character could be accepted to encode separately at last, these four are also needed to remove, unless any experts could provide the evidence cited from them. This behavior is useful for the future.

I would be interested in knowing how common U+25997 𥦗 is. The source is is GHZ-42731.08, but 《汉语大字典》（第二版） p. 2922 does not give any sources, and merely states 同“窓（窗）”.



As L F Cheng says (#14310), maybe the dictionary form is a normalization of ⿱宀总. If 𥦗 is not a common form, then China can consider changing the glyph for U+25997 to ⿱宀总.

Response to Comment #14310.
The G-Source reference for U+25997 𥦗 is GHZ-42731.08, that means it is cited from 《汉语大字典》, see below.



In fact, 𥦗 and ⿱宀总 are both needed for China, so I think the normalization will affect so many databases. Unifying doesn’t mean we throw the character/form away. Dr. Lu has explained this issue again and again at the IRG meeting. What we need to consider should be as below.

1) Is there at least one piece of evidence shows 𥦗 and ⿱宀总 is not the variant of 窗? If yes, the character must be encoded separately, or we need to consider other unification; if not, ⿱宀总 should be unified with 𥦗, and consider how to update UCV.

2) The standard must go faster than reality. Macao SAR has registered several IVSes, but not all the IVSes has been run well in the OSes, even if in macOS, iOS, iPadOS and so on. We can not say this is not meaningful and ask Macao SAR not register IVS in future.

The most useful evidence to encode this character separately is the surname usage. 《中华姓氏源流大辞典》 shows the surname is cited from 《祁东县志》 and shows the reading and it is the variant of 窗, but the evidence cited from 《祁东县志》 just shows the character. If there is any evidence in 《祁东县志》 shows this character is not related to 窗, this character must be disunified.

Response to Comment #14306 and #14318.

There are more than 10,000 entries related to current U+25997 𥦗 in my databases. The form for U+25997 𥦗 is also stable. And so many databases had been used this form. It is better to keep the code chart stable and not normalize the G glyph.

[总] may be a special case for identifying abstract shape, because it came from simplification in {*TSOŊ (悤)} field.

The two PRC simplified forms that contain [总] is [总] from [總] and [聪] from [聰]. Another over-simplified forms, [𩨂] (GFZ) from [驄], [𮴨] (GZH) from [璁], [𱶚] (Leizhou) from [𥡥], establish strong correlation between [总] and [悤]. So if a true ⿱[宀][总] exists, it may connect to ⿱[宀][悤], that is WS2021-01052.

01052

宀 40.11.3

TD-4A59

TS 14 · IDS ⿱宀悤

Evidence accepted, IRG 57.

[窗] is related to [窻] [牕] [窓], M.C. status 初 initial 江 final 平 tone, in phonetic symbol {*TSOŊ (悤)}. [總] in 精 initial 東 final 上 tone, [聰] is related to [聦], in 清 initial 東 final 平 tone, all in phonetic symbol {*TSOŊ (悤)}. So the potential existed ⿱[宀][总] may be also in {*TSOŊ (悤)}.

Reply to #14310,

Radical 穴 comes from cognition. Seal [窗] should be analysed as ⿱[穴][囱], and seal [窻] should be analysed as ⿱[穴][悤 (phonetic)].
Then [悤] becomes |怱| |忽| |𱝼| |𢗀|. The |八| from |穴| is regarded to form a unity with |𱝼| |𢗀|, and thus constitutes the |忩| |总|.

Reply to #14318. I searched 𥦗 in 瀚堂典藏数据库(HYTUNG BOOKS), there are 13 pages of the result.

Reply to #14319.
Again, to clarify, I am suggesting encoding the character seperately NOT JUST AND NOT MAINLY because some information processing systems don't support IVD. It is not about unifying a character meaning throwing away a character. So please do not involve unnecessary things into this particular case.

Wang Xieyang's argument on Oct 16 seems to be summarized in two points with regards to unification rules.

1. This character has developed distinct identity from 𥦗 in some regions (namely China).
2. This character has different underlying structure with 𥦗, signified by the orientation of two dots in the middle.

The criterion #1 is about subjective judgement, so we will defer it to experts from China, but it is perhaps only applicable to post-簡化字 usage, where 总 is formulated as a canonical shape.

As for #2, we have seen some analogous case in this WS:

01842

木 75.12.2

SAT-06594

TS 16 · IDS ⿱敞木

Not unified to 𢿵 U+22FF5, different structures, IRG 60.

, where it can be justified if we can prove that this character is an outcome of separate development path from 𥦗. In favor of the assertion, as Kushim pointed (comment #14323), we also have in this WS

01052

宀 40.11.3

TD-4A59

TS 14 · IDS ⿱宀悤

Evidence accepted, IRG 57.

that looks like its traditional variant without the 穴 radical. On the other hand, we also notice that, in Japanese 新字体, "window" is 窓 while 總 becomes 総, which suggests that the difference might be irrelevant. Do we have an evidence that shows somebody clearly recognizes the character as composition of 宀 and 总?

Change Radical to 40.0 (宀), SC=9, FS=4

This ideograph was present in the first and second drafts of GB 18030-2022 Amendment 1, specifically in the Plane 10 repertoire at U+A015F and U+A0271, respectively. It was also included in the first draft of the CJK Unified Ideographs Extension I repertoire at U+2EC5E. At the end of June, per feedback from China, nearly 50 ideographs were removed from the Extension I repertoire, almost all of which were in this working set. One can reasonably conclude that China expected them to remain in this working set, otherwise they would not have been removed. If China had not requested that those ideographs be removed, this particular ideograph would already be encoded in Extension I, and this entire discussion would be moot.

Given the sheer amount of discussion, it would be prudent for the IRG to err on the side of disunification in this particular case. What I wrote in the previous paragraph can also not be ignored. Let us accept this ideograph and move on to more important issues in this working set.

Extension I is formmaly approved by WG2 according to WG2 5234.

https://www.unicode.org/wg2/docs/n5234-WG2-M70-Recommendations.pdf
The character ⿱宀总 is in the first and second draft of GB 18030-2022 Amendment 1 and then is still included in the draft of Extension I after China removes 55 characters out of the set. It is included in Extension I in the last miniute. However, experts from China found that ⿱宀总 was in IRG WS2021. Seeing the character not suggested to be unified or withdrawn, in order to reduce the possibility of encoding duplicates, China decided to move this character out from Extension I. It will be very disappointing and reluctant if the character is unified. In fact, this character was formally approved by WG2.
What's more, it is reasonable to think that it is better to encode the character seperately as I commented in comment #14307.
The character is already in the PUA fonts of The Natural Resources Surveying and Monitoring Institute of Jiangxi Province(江西省自然资源测绘与监测院), The Ministry of Public Security of the People’s Republic of China(中华人民共和国公安部) and Zhonghua Book Store(中华书局). Unifying this character will simply cause the abuse of PUA because noramlly, they will be recognized as different characters with same meaning by people.
Last but not least, please try to think about this: If ⿱宀总 is unified to 𥦗 (U+25997) , then the radical of ⿱宀总 should be 穴, which is also reluctant.
I sincerely suggest to encode this character seperately based on what I wrote in this comment.

In principle I think ⿱宀总 could be unified with 𥦗 (U+25997), and represented using IVS. However, I think there is merit in Wang Xieyang's argument that ⿱宀总 and ⿱穴𢗀 are intrinsically different characters because they have different radicals. Ken's argument (#14618) also makes me inclined to take a pragmatic view of the encoding of this character. Therefore I now support encoding, and as sufficient evidence has now been supplied, GDM-00313 should be moved back to the M-set.

The following picture shows ⿳宀丷工 which should be unified with 空. I once mentioned this form in Comment #14286.


招子庸: 《粤謳》, 道光戊子年本, folio 1B

⿱宀总 is mentioned in _Long Story of Short Forms_ by Roar Bökset (https://vitterhetsakademien.bokorder.se/en-us/article/4450/long-story-of-short-forms), under the 窓 entry.