Chinese Character Code for Information Interchange

Lore & Background

CCCII was designed as a 94n set per ISO/IEC 2022, with each Chinese character represented by a 3-byte code. Its maximum theoretical capacity is 830,584 characters, though practical encoding is less due to reserved spaces for variant characters. The 94 ISO 2022 planes are grouped into 16 layers, with layer 1 containing non-hanzi and hanzi, layer 2 simplified characters, and layers 3–12 further variants. Layers 13 and 14 add Japanese and Korean support, while layer 16 holds other characters. A variant of an earlier version, the East Asian Character Code (EACC), is used by the Library of Congress as part of MARC-8.

Reader's Guide

CCCII's significance lies in its early and sophisticated encoding of traditional Chinese characters, including a systematic approach to variant forms. It was a direct precursor to Unicode: work at Apple based on the Research Libraries Group's CJK Thesaurus, used to maintain EACC, was incorporated into Unicode's Unihan set. Despite its innovative design, adoption was limited to libraries in the United States, Hong Kong, and Taiwan due to specialized hardware requirements and difficulty in determining root versus variant characters. As of 2009, EACC remained in extensive use for bibliographic purposes. The structure of CCCII has been both praised for its thoughtfulness and criticized for its fixed, one-dimensional approach to variant relationships. Unicode hanzi characters are referenced to CCCII and EACC codes in the Unihan database, though mapping is not complete due to differing unification criteria.

Did You Know?

• CCCII predates Big5 (1984) and CNS 11643 (1986).
• EACC contains fewer characters than the most recent versions of CCCII.
• CCCII defines roughly 53940 code points as of its 1987 edition.
• The code 0x212320 is used by some implementations as an ideographic space.

Origins and Structural Architecture

Developed by the Chinese Character Analysis Group in Taiwan, CCCII first appeared in 1980 and saw major expansions in 1982 and 1987. It stands as one of the earliest and most technically refined encodings for traditional Chinese, arriving before both Big5 in 1984 and CNS 11643 in 1986. The system is built on the ISO/IEC 2022 94n framework, meaning every character is encoded as a three-byte sequence where each byte is a 7-bit value falling between 0x21 and 0x7E. This yields a theoretical ceiling of 94 cubed, or 830,584, code points, though the actual repertoire is smaller because variant glyphs consume related planes. The 94 planes are organized into sixteen layers of six planes apiece, with the final layer holding only four. Layer 1 carries non-hanzi symbols and the most frequently used hanzi; layers 2 through 12 accommodate simplified forms and additional variants; layers 13 and 14 serve Japanese kana and Korean hangul respectively; layer 15 is reserved; and layer 16 holds miscellaneous characters. The 1987 edition defined roughly 53,940 code points, and a 1989 draft pushed that figure to 75,684.

The Layered Variant Character System

CCCII's most distinctive innovation is its treatment of character variants through a homologous layering scheme. Layer 1 holds the primary traditional Chinese forms, while layer 2 places their simplified counterparts at identical row and cell coordinates. Layers 3 through 12 extend the same positional alignment to further variant shapes, so a single conceptual character can be addressed across multiple layers at the same grid position. This multi-dimensional mapping was praised by Ken Lunde, who called it one of the most well thought-out character set standards from Taiwan and said its structure was to be truly admired, though he noted that OpenType variant form substitution could deliver comparable functionality. Christian Wittern of Hanazono University offered a sharper critique, arguing that the relationships among character variants are too intricate to be captured in a fixed, one-dimensional, hard-wired codetable. The scale of the challenge is evident in the 1989 draft, which listed 44,167 unique characters alongside 31,517 variant entries, underscoring just how deeply the layered architecture had to penetrate the complexity of Chinese script.

Adoption, Limitations, and the Library Domain

Despite its technical depth, CCCII never achieved broad consumer adoption. Its primary domain was library and bibliographic systems. By 1995, the encoding was in use mainly in libraries across the United States, Hong Kong, and Taiwan. A trimmed variant, EACC (East Asian Character Code, designated ANSI/NISO Z39.64), was folded into the Library of Congress's MARC-8 framework, where it was assigned the private-use F-byte 0x31 under the ANSI X3.41 implementation of ISO 2022. EACC contained only 15,686 characters, a modest subset of CCCII's full repertoire. Several practical obstacles constrained wider uptake: reliance on specialized hardware, the difficulty of deciding when a root form versus a variant should be displayed, and the absence of firmly established reference glyphs to guide those decisions. Outside the library world, Big5 became the dominant encoding for Chinese in those territories, especially before Unicode gained traction. Even as late as 2009, EACC remained in extensive bibliographic use, and the Library of Congress published mapping tables linking EACC to Unicode for hanzi, hangul, kana, and punctuation.

Precursor to Unicode and Enduring Cross-Reference Legacy

CCCII's influence on modern character encoding extends far beyond its own operational life. Its most tangible legacy is embedded in Unicode's Unihan database. At Apple, engineers built a CJK character cross-reference database drawing on the Research Libraries Group's CJK Thesaurus, which had been used to maintain EACC. That work was directly incorporated into the development of Unicode's Unihan set. Within Unihan, individual hanzi carry reference keys kCCCII and kEACC that point back to their corresponding codes in those earlier systems, preserving a traceable link to the 1980s encoding. However, because Unicode's unification criteria—shaped by Japanese JIS X 0208 and by the Association for a Common Chinese Code in China—differ from CCCII's own variant-handling philosophy, not every CCCII variant maps to a distinct Unicode code point. CCCII thus occupies a foundational position in the history of CJK encoding, its layered variant logic and cross-reference data having seeded structures that continue to inform how Chinese characters are organized and referenced in today's universal character set.