Mojikyō

Mojikyō (Japanese: 文字鏡), also known by its full name Konjaku Mojikyō (今昔文字鏡, lit. '(the) past and present character mirror'), is a character encoding scheme. The Mojikyō Institute (文字鏡研究会, Mojikyō Kenkyūkai), which published the character set, also published computer software and TrueType fonts to accompany it. The Mojikyō Institute, chaired by Tadahisa Ishikawa (石川忠久),^[1] originally had its character set and related software and data redistributed on CD-ROMs sold in Kinokuniya stores.^[2]

Mojikyō
Konjaku Mojikyō
今昔文字鏡

The Mojikyō character map highlighting the Taiwanese kana　セ^[note 1]
Developer(s)	Tadahisa Ishikawa (石川忠久) Tokio Furuya (古家時雄) Mojikyō Institute (文字鏡研究会)
Initial release	1.0 / July 1997; 25 years ago (1997-07)
Final release	4.0 / December 15, 2018; 4 years ago (2018-12-15)
Operating system	Microsoft Windows
Size	51MB
Available in	Japanese
Type	Character set bundled with fonts and a character map
License	Proprietary
Website	mojikyo.org

Conceptualized in 1996,^[3] the first version of the CD-ROM was released in July 1997.^[4] For a time, the Mojikyō Institute also offered a web subscription, termed "Mojikyō WEB" (文字鏡WEB), which had more up-to-date characters.^[5]

As of September 2006^[update], Mojikyō encoded 174,975 characters.^[6] Among those, 150,366 characters (${\displaystyle \approx }$86%) then belonged to the extended Chinese–Japanese–Korean–Vietnamese (CJKV)^{[note 2]} family.^[5] Many of Mojikyō's characters are considered obsolete or obscure, and are not encoded by any other character set, including the most widely used international text encoding standard, Unicode.

Originally a paid proprietary software product, as of 2015, the Mojikyō Institute began to upload its latest releases to Internet Archive as freeware,^[7] as a memorial to honor one of its developers, Tokio Furuya (古家時雄), who died that year.^[3] On December 15, 2018, version 4.0 was released. The next day, Ishikawa announced that without Furuya this would be the final release of Mojikyō.^[3]

Premise

The Mojikyō encoding was created to provide a complete index of Chinese, Korean, and Japanese characters. It also encodes a large number of characters in ancient scripts, such as the oracle bone script, the seal script, and Sanskrit (Siddhaṃ). For many characters, it is the only character encoding to encode them, and its data is often used as a starting point for Unicode proposals.^[8][9] However, Mojikyō has much looser standards than Unicode for encoding, which leads Mojikyō to have many encoded glyphs of dubious, or even unintentionally fictional, origin.^[10][11] As such, while many non-Unicode Mojikyō characters are suitable for addition to Unicode, not all can become Unicode characters, due to the differing standards of evidence required by each.

Composition

The Mojikyō fonts (文字鏡フォント) are TrueType fonts that come in a ZIP file and are each around 2–5 megabytes; the different fonts contain different numbers of characters.^{[note 3]} Also included is a Windows executable that implements a graphical character map, the "Mojikyō Character Map" (文字鏡MAP), MOCHRMAP.EXE.^{[note 4][note 5]} MOCHRMAP.EXE allows users to browse through the Mojikyō fonts, and copy and paste characters in lieu of typing them on the keyboard. As opposed to the regular Windows character map, or for that matter KCharSelect, which both support TrueType fonts, MOCHRMAP.EXE displays the numbered Mojikyō encoding slot of the requested character.^{[12][note 6]} In order for MOCHRMAP.EXE to work, all Mojikyō fonts must be installed.^{[note 7]}

Encoding

When referring to a character encoded in Mojikyō, the format MJXXXXXX is often used, similar to the U+XXXX format used for Unicode. For example, hentaigana U+1B008 𛀈 has Mojikyō encoding MJ090007 and Unicode encoding U+1B008.^[13] A difference, however, is that Mojikyō encodings displayed this way are decimal, while Unicode's U+ encoding is hexadecimal.

From the earliest days of Unicode, Mojikyō has both influenced—and been influenced by—the standard. Glyphs originating from Mojikyō first appear in a proposal to the Ideographic Rapporteur Group (IRG),^{[note 8]} which is responsible for maintaining all CJK blocks in Unicode,^[14][15] on 18 April 2002.^[16] In May 2007, Mojikyō played a minor role in an eventually successful series of proposals to encode the Tangut script in Unicode;^{[17][note 9]} Mojikyō already had within its encoding 6,000 Tangut characters by October 2002.^[6]

The Unicode Standard's Unihan Database refers to Mojikyō as the "Japanese KOKUJI Collection" (日本国字集),^[18] abbreviated "JK".^[19][20] For example, U+2B679 𫙹 ,^{[note 10]} an ideograph read in Japanese as burizādo (ブリザード, lit. 'blizzard'), has a J-Source^{[note 11]} equal to JK-66038. All Unicode characters with a JK-prefixed J-Source originate from Mojikyō.^{[21][note 12]} According to Ken Lunde, a subject matter expert in character encodings and East Asian languages, as of Unicode 13.0, 782 ideographs in Unicode originate from Mojikyō, split somewhat evenly between two blocks: CJK Unified Ideographs Extension C, with 367, and CJK Unified Ideographs Extension E, with 415.^[20][22] Not all Unicode characters with Mojikyō origins (JK-prefixed J-Sources) have the same representative glyph in the code chart as in the Mojikyō font;^{[note 13]} some characters had their shapes changed before final encoding, as investigation showed the shapes assigned by the Mojikyō Institute were wrong.^{[11][note 14]}

Blocks

As of September 2006^[update] it encoded 174,975 characters.^[6] Among those, 150,366 characters then belonged to the extended CJKV^{[note 2]} family.^[5] Many of the encoded characters are considered obsolete or otherwise obscure, and are not encoded by any other character set, including the international standard, Unicode. Each Mojikyō character has a unique number, and the characters are organized into blocks.

Mojikyō puts CJKV characters in different blocks according to their traditional Kangxi radical. Common radicals containing an especially high number of characters, such as Radicals 9 (人) and 162 (⻌), are split further by stroke order.^{[note 15]}

No unification

Unlike Unicode, Mojikyō purposely avoids Han unification; no attempt at compactness of the encoding is made, nor is there an attempt to keep all common characters below U+FFFF as there is in Unicode.

Unicode, on the other hand, sorts its CJK into blocks based on how common they are: the most common are generally put into the Basic Multilingual Plane,^{[note 14]} while those that are rare or obscure are put into the Astral Planes.

For example, Radical 9 has two characters where Unicode has one: MJ054435 (令), and MJ059031 (令), both represented in Unicode as U+4EE4 令 .

License

Mojikyō is proprietary software under a restrictive license. Originally, the Mojikyō Institute tried to prevent its character data from being used, and threatened those who published conversion tables to and from its character set. In July 2010, the Mojikyō Institute abandoned its legal efforts to stop at least one Japanese user from publishing conversion tables or converting characters encoded in Mojikyō to Unicode or other character sets.^[23] Mere data, sometimes including the shapes of letters, are considered in many jurisdictions to be common property as they do not meet the threshold of originality.^{[note 16]}

Due to this legacy, however, GlyphWiki [ja] disallowed Mojikyō data as of 2020.^[24]

Collected writing systems

Living

• Chinese — Hanzi
• Japanese — Kanji, Kana (including Hentaigana)
• Korean — Hanja
• Latin alphabet with diacritics
• Cyrillic script with diacritics

Dead or obsolete

• Ancient Chinese
  • Oracle bone script
  • Seal script
• Taiwanese kana
• Vietnamese — Chữ Nôm
• Sanskrit — Siddhaṃ
• Tangut script
• Sui script

See also

• Han unification
• JIS X 0208
• List of CJK fonts
• TRON

References

1. [About Mojikyō]. Mojikyō Institute (in Japanese). Archived from the original on 3 February 2001. Retrieved 6 July 2020.
2. [Welcome to the world of Mojikyō!] (in Japanese). Kinokuniya KK. Archived from the original on 4 March 2005. Retrieved 5 July 2020.
3. Ishikawa, Tadahisa (August 2015). "古家時雄君を悼む" [Tokio Furuya, we grieve your death]. Mojikyō Institute (in Japanese). Retrieved 8 July 2020.
4. Konjaku Mojikyō 今昔文字鏡 (in Japanese), July 1997, ISBN 9784314900034
5. [About Mojikyo] (in Japanese). Kinokuniya KK. Archived from the original on 27 April 2010. Retrieved 5 July 2020.
6. [What is Mojikyō?] (in Japanese). Kinokuniya KK. Archived from the original on 5 February 2005. Retrieved 5 July 2020.
7. "Search: creator:"MOJIKYO Institute"". Internet Archive. Retrieved 6 July 2020.
8. Takada, Tomokazu; Yada, Tsutomu; Saito, Tatsuya (18 September 2015). Proposal for hentaigana (PDF). Translated by Kobayashi, Tatsuo; Kobayashi, Daniel. Information Processing Society of Japan. L2/15-239. Retrieved 5 July 2020 – via Unicode Consortium.
9. Hiura, Hideki; Kobayashi, Tatsuo; et al. (31 October 2003). Ideograph Variation Selector and Variation Collection Identifier. Open Internationalization Initiative. L2/03-413. Retrieved 5 July 2020 – via Unicode Consortium.
10. Takada, Tomokazu [高田智和]; Oda, Tetsuji [織田哲治]; et al. (26 August 2013). 平成25年度第3回文字情報検討サブワーキンググループ議事録 [Meeting Minutes of the Third Character Information Examination Sub-Working Group of 2013 (Heisei 25)] (PDF). Information Technology Promotion Agency, Government of Japan (in Japanese). p. 2. Retrieved 6 July 2020. 文字鏡研究会の関係者にヒアリングしたところ、オランダから提案されたWG2 N36981には文字鏡のフォントが使用されているが、文字鏡研究会は関与しておらず、提案内容についても疑問があるとのことであった。[According to an interview with a representative of the Mojikyō Institute, a Mojikyō font is used in WG2 N36981 proposed by the Netherlands, but the Mojikyō Institute itself is not involved with the proposal; it furthermore has doubts about some of the content of that proposal.]
11. Suzuki, Toshiya [鈴木俊哉] (30 July 2009). [Investigation on Glyphs collected from "Index to Collection of Inscriptions of the Yin-Zhou Period" to submit to CJK Unified Ideographs]. IPSJ SIG Technical Report (in Japanese). Information Processing Society of Japan. 2009-DD-72 (7): 2 – via Internet Archive. しかし、拡張Cの標準化作業が8年の長期にわたり、また事後的に用例が必須とされたため、正式に公布された拡張C漢字の典拠は当初の典拠とはかなり異なるものとなっている。たとえば日本では当初は文字鏡研究会によって選定された1000文字程度の漢字を申請していた[。] [...] 典拠用例確認は文字鏡とは独立に行なわれたため、字形が文字鏡漢字から変更されたものも多い。[As the standardization effort for CJK Unified Ideographs Extension C has been eight long years in the making and examples of kanji have been requested after their encoding, the officially promulgated Extension C kanji standard is quite different from the original standard. For example, we, the Government of Japan, initially applied for about 1,000 kanji selected by the Mojikyō Institute[.] [...] Since the verification of the kanji was performed independently of the Mojikyō Institute, the character shapes were often changed from Mojikyō's version of that same codepoint.]
12. Ishikawa, Tadahisa (25 May 1999). "パソコン悠悠漢字術 今昔文字鏡徹底活用" [Kanji on your PC, Made Easy—The Complete Mojikyō Manual]. Mojikyō Institute. Retrieved 6 July 2020.
13. MJ文字情報一覧表 [Table of MJ Character Encodings] (in Japanese). Information Technology Promotion Agency. Retrieved 5 July 2020.
14. "Unicode Standard Annex #45: U-source Ideographs". The Unicode Standard. Unicode Consortium.
15. "Appendix E: Han Unification History" (PDF). The Unicode Standard. Unicode Consortium. March 2020.
16. "CJK Extension C1 From Japan". Ideographic Rapporteur Group. IRG#19 N895 – via The Chinese University of Hong Kong's Department of Computer Science and Engineering. N895-Japan_C1
17. Cook, Richard (9 May 2007). Proposal to encode Tangut characters in UCS Plane 1 (PDF). UC Berkeley Script Encoding Initiative. p. 4. L2/07-143 – via Unicode Consortium.
18. Jenkins, John H.; Cook, Richard; Lunde, Ken, eds. (5 March 2020), "kIRG JSource", Unicode Standard Annex #38, Unicode Consortium
19. Kobayashi, Tatsuo (3 December 2001). "List of Japanese Ideographs which may be proposed in Extension-C". ISO/IEC JTC1/SC2/WG2/IRG N853.
20. Ken Lunde [@ken_lunde] (6 July 2020). "In particular, all 782 JK-prefixed ideographs are indeed from 今昔文字鏡 per IRG N862. Most were encoded in #ExtensionC, and the stragglers were encoded in #ExtensionE." (Tweet). Retrieved 6 July 2020 – via Twitter.
21. Ken Lunde [@ken_lunde] (6 July 2020). "JK-prefixed J-Source ideographs came from 今昔文字鏡, which are in Extensions C and E (the mention of Extension D was simply that what became Extension E was originally targeted to become Extension D)" (Tweet). Retrieved 6 July 2020 – via Twitter.{{cite web}}: CS1 maint: url-status (link)
22. Ken Lunde [@ken_lunde] (6 July 2020). "367 JK-prefixed ideographs are in Extension C, and the remaining 415 are in Extension E." (Tweet). Retrieved 6 July 2020 – via Twitter.
23. "終戦宣言" [Announcement: The War is Over]. 青蛙亭漢語塾 [Seiwatei's Kanji Cram School] (in Japanese) (28 January 2016 ed.). 21 July 2010. Retrieved 7 July 2020.
24. "データ・記事のライセンス" [License of our data and articles]. GlyphWiki (9 June 2010 ed.). Retrieved 6 July 2020. 今昔文字鏡およびその関連製品、データは、そのライセンス上グリフウィキには用いることができません。文字鏡番号（独自部分）および文字鏡のフォントに収録されているグリフそのもの、およびそれを参照、利用して作成していると判断できる情報は、グリフウィキに登録する際の典拠とすることはできませんので、ご協力をお願いいたします。 [Konjaku Mojikyō and related products and associated data are licensed in such a way that they are incompatible with our above GlyphWiki license. Neither the number of the Mojikyō encoding slot, nor the appearance of the glyph itself in Mojikyō's fonts, nor any information whatsoever that can be judged to have been gathered by referring to a Mojikyō product, can be used when entering data into GlyphWiki. We absolutely cannot accept Mojikyō data. Please cooperate with us.]

Notes

1. As yet, lacks a Unicode encoding, so is approximated here with CSS and U+30BB セ KATAKANA LETTER SE.
2. For Korean, Hanja are referred to. For Vietnamese, Chữ Nôm.
3. Download the file MojikyoCmap400ALL49TTF.7z from the official website
4. English name from the title of the window produced by running the executable; Japanese name from the icon of the executable.
5. Also called the "Mojikyō Cmap".
6. See the screenshots on the official website
7. Into the system fonts directory C:\Windows\Fonts.
8. As of 2019, the IRG rebranded as the Ideographic Research Group.
9. The history of the encoding of the Tangut script is quite complicated, see Tangut (Unicode block) § History for a full listing of all the related proposals and a timeline.
10. Ideographic Description Sequence: ⿰魚嵐
11. This is a column name in the Unihan database; ⟨J⟩ here is short for "Japanese glyph source". The full name of the column is kIRG_JSource. Under Han unification, there are nine such sources. See §3.1 of UAX#38 for a complete list and more information.
12. Other J-Source prefixes exist, such as J4, meaning the character originates from JIS X 0213:2004.
13. That is to say, a glyph made up of the same radicals in the same positions.
14. Errors in large collections of ideographs are, of course, not uncommon. Such errors even accidentally occur in well funded government-produced collections, such as the famous kanji from unknown sources in the Japanese Industrial Standards Committee's JIS X 0208 double-byte character encoding standard. All of these JIS X 0208 error kanji (幽霊文字; e.g., 彁) have made their way into Unicode despite not being "real" kanji.
15. For proof, see the list in the Mojikyō Character Map, MOCHRMAP.EXE.
16. See also: fictitious entry; trap street.

External links

• Official website