Sources & Credits

The Chopmark Lookup tool, and the character and script pages that go with it, rely on character data, fonts, and software from a number of open projects. Thank you to everyone who has made their work freely available.

Character data

• Unihan Database (© Unicode, Inc.) — pinyin readings and English definitions
• Unicode Character Database (© Unicode, Inc.) — the Equivalent_Unified_Ideograph property, which lets Lookup by Radical recognise a radical copied from a dictionary chart (⺅) as the form that appears inside characters (亻)
• KANJIDIC2 — the Electronic Dictionary Research and Development Group (Jim Breen / Monash University) — Japanese on'yomi and kun'yomi readings, English meanings, and jōyō / jinmeiyō classifications used by the Japanese library. Used under the Creative Commons Attribution-ShareAlike 4.0 licence.
• CC-CEDICT (published by MDBG) — English definitions for rare characters the Unihan database does not gloss. Used under the Creative Commons Attribution-ShareAlike 4.0 licence.
• Wiktionary — English definitions for rare characters, extracted with wiktextract (Tatu Ylonen) via kaikki.org. Its Chinese radical appendix also supplies the character lists, grouped by additional stroke count, on the Characters by Radical pages. Used under the Creative Commons Attribution-ShareAlike 4.0 licence.
• 康熙字典 (Kangxi Dictionary), 1716 — used only where it records that one character is simply another character's variant form, which lets a rare character inherit its parent's meaning. The dictionary itself is in the public domain. The digital text is the kangxiDictText edition (MIT licence), which builds on the earlier transcription by Simon Wiles.
• zi.tools, the character aggregator: for the rare characters no open dictionary describes, its pages supplied Mandarin readings, the relations between a variant form and its standard character, and the Chinese glosses from which the English meanings on this site were written afresh. Consulted once for each character, slowly, and stored here so that nothing is fetched from it at runtime. The table of component forms that the radical tool treats as one part (壬 and 王, 戶 and 户) was drafted from the same author's structured copy of the IRG's Unifiable Component Variants list, iwds.
• CNS 11643 全字庫, Taiwan's Ministry of Digital Affairs, a second time: the 17,703 characters the national standard records that Unicode does not, most of them name-forms from Taiwan's household, land and tax registers, are in the Expanded library with the standard's own readings and component sequences, each named by its CNS code, and about a thousand characters Unicode encoded from other countries' submissions carry the standard's reading and code as well. The same component sequences make the newest Unicode block, Extension J, searchable by radical. Released under the Open Government Data License, version 1.0.

Character decomposition

The Lookup by Radical tool works out which characters contain which parts from two freely-shared Ideographic Description Sequence databases, used together so a part is found whether it is listed by visual shape or by structural analysis.

• BabelStone IDS, maintained by Andrew West (魏安) — a visual-first decomposition of the CJK ideographs, released for any personal or commercial use without restriction
• cjkvi-ids (CJKVI Database), derived from the CHISE IDS database — used under the GNU General Public License, version 2

Modern Hanzi & Japanese glyphs

• Noto Sans CJK (Google) — the SC variant for Hanzi forms and the JP variant for Japanese kanji & kana forms; subsets of the SC variant are also served to your browser, on every page of this site, so that rare characters display properly wherever they appear (subsets of Plangothic and of TW-Sung-Plus, below, travel with them for the characters Noto lacks)
• Noto Sans Mongolian (Google) — the upright vertical glyphs for Manchu-script chopmarks, shipped as a webfont
• Plangothic — the rare and supplementary-plane CJK characters (Extensions B–J, from its P1 and P2 fonts) in the Chopmark Lookup tool, both for matching and, as webfont subsets, for display
• LXGW WenKai — a kaiti (brush-influenced regular script) face, used behind the scenes to give the matcher a second, more handwritten template of simple characters, which is closer to the shape most people draw. Build-time only, never served to your browser.
• TW-Sung-Plus and TW-Kai-Plus, the CNS 11643 全字庫 fonts from Taiwan's Ministry of Digital Affairs: the only faces that cover the characters the standard records and Unicode does not. The Sung face draws those characters in the Expanded library and is served to your browser as webfont subsets wherever one of them appears; the Kai face gives the matcher its handwritten second template for them, build-time only. Released under the Open Government Data License, version 1.0.

The four families above them are distributed under the SIL Open Font License.

Seal, Bronze, Oracle, and Liushutong script forms

• hanziyuan.net — Richard Sears' "Uncle Hanzi" project, a remarkable freely-shared paleographic character database
• 崇羲篆體 (ChongXi Small Seal), Academia Sinica's 小學堂 — a calligrapher's hand for the Shuowen small-seal forms. Used under the Creative Commons Attribution-NoDerivatives 3.0 Taiwan licence.
• 全字庫說文解字 (CNS11643 Shuowen Jiezi), Taiwan's Ministry of Digital Affairs — a second small-seal source, digitized from a woodblock edition, so most characters can be shown in two different engraving hands. Used under the Creative Commons Attribution-NoDerivatives 3.0 Taiwan licence.
• CNS 11643 全字庫 character properties, Taiwan's Ministry of Digital Affairs. The national standard's own readings, which supply a Mandarin pronunciation for tens of thousands of rare characters that no dictionary we hold reads. Released under the Open Government Data License, version 1.0.
• Academia Sinica's 小學堂 (xiaoxue) character query interface — Chu bamboo-slip (楚系簡帛) and transmitted ancient (傳抄古文字) forms, which are the only ancient attestation on file for a number of characters, including several chopmarks in this collection. Released by Academia Sinica into the public domain under CC0 1.0.
• Wikimedia Commons contributors — additional ancient-script SVGs

Where these forms come from a font, that font is used only to draw the glyph images shown on the page. The font files themselves are never altered and never redistributed.

Calligraphy script forms (running, grass, and clerical, shown in the Direct Lookup tool, the Chopmark Library, the Script Style Reference and Characters by Radical pages, and alongside chopmark photographs on the coin pages)

• Zhi Mang Xing (Wei Zhimang) and Long Cang (Chen Xiaomin): running script, via Google Fonts under the SIL Open Font License
• Liu Jian Mao Cao (Liu Zhengjiang): grass script, Google Fonts (SIL Open Font License)
• 青柳隷書 (Aoyagi Reisho) and 衡山毛筆フォント行書 / 草書 (Kouzan Gyōsho and Sōsho), by calligrapher Aoyagi Kouzan (digitized by Shimo, 武蔵システム): clerical, running, and grass forms, free for any use
• 草书 (Caoshu), a wild-cursive (狂草) font by mikhailcai modeled on the Tang master Sun Guoting (孙过庭): grass script; the author has waived copyright and it is free for any use
• 书体坊于右任标准草书 (Yu Youren Standard Cursive), digitized by 书体坊 / Founder after the modern master Yu Youren (于右任): his standardized, legible cursive system (标准草书), used here only as rendered glyph images
• HanWang LiSu (王漢宗中隸書), the free fonts of Prof. Hann-Tzong Wang: clerical script (GPL)
• UnYetgul (은예글), Un Koaunghi / Un Fonts: clerical hanja (GPL)

These calligraphy fonts are used only as rendered glyph images on the page, never redistributed as font files.

Build tools

• fontkit — glyph outline extraction
• @napi-rs/canvas — server-side image rasterization
• subset-font — webfont subsetting
• node-potrace — tracing scanned glyph images to clean outlines, a JavaScript port of Peter Selinger's Potrace, used under the GNU General Public License, version 2