Code page 952
| Company type | Japanese Industrial Standard (JIS) |
|---|---|
| Industry | Character encoding standard |
| Predecessor | JIS C 6226 (initial name) |
| Founded | 1978 |
| Headquarters | Japan |
| Products | Two-byte character set of 6,879 graphic symbols for Japanese text encoding |
JIS X 0208 is a two-byte character set established as a Japanese Industrial Standard, encompassing 6,879 graphic symbols utilized for representing text, geographic names, personal names, and other elements in the Japanese language. Its formal designation is . Initially introduced as JIS C 6226 in 1978, this standard underwent updates in 1983, 1990, and 1997. IBM identifies it as Code page 952, while the original 1978 edition is known as Code page 955.
Usage scope and interoperability
The primary function of JIS X 0208 is to facilitate communication between data processing systems and related devices, enabling the handling of textual data.
Partial adoption of the standard is regarded as incompatible. The early versions divided characters into level 1 and level 2 categories, with certain variant characters reallocated between these levels during subsequent revisions. Some platforms, such as the early NEC PC-9801, implemented only level 1 or exclusively non-kanji characters; however, these limited implementations were never officially recognized as compliant.
Although the 1997 revision included measures to improve compatibility, JIS X 0208 generally lacks certification for self-compatibility or official manufacturing endorsement. As a result, no products have been formally declared JIS X 0208-"compatible." While terms like and appear within the standard documentation, their interpretation varies among users and lacks uniformity.
Code charts
Lead byte
The initial byte in the encoding corresponds to the row or cell number plus hexadecimal 0x20 (decimal 32). For instance, a code beginning with 0x21 indicates row number 1, with cell 1 continuing at byte 0x21 (decimal 33).
Lead bytes representing non-kanji characters link to specific charts on this page that list the encoded characters. Those corresponding to kanji direct users to Wiktionary’s kanji index sections.
(Note: Subsequent rows continue similarly featuring Kanji Level 1 and Level 2 entries up to row number eighty-four.)
Character Set 0x21 (Row 1, Special Symbols)
Different manufacturers apply varying Unicode assignments to this character group. For example, Microsoft links kuten 1-29 (JIS 0x213D) to U+2015 (Horizontal Bar), whereas Apple designates it as U+2014 (Em Dash). Likewise, Microsoft maps kuten 1-61 (JIS 0x215D) to U+FF0D (Fullwidth Hyphen-Minus), in contrast to Apple’s use of U+2212 (Minus Sign). The wave dash's Unicode representation also differs depending on the vendor.
When paired with JIS X 0208 in encoding schemes such as Shift JIS, EUC-JP, or ISO 2022-JP, ASCII and JISCII punctuation marks—highlighted in yellow—may be mapped into the Halfwidth and Fullwidth Forms Unicode block.
... [table continues with similar entries, reduced for brevity] ...Character Set 0x22 (Row 2, Special Symbols)
The majority of characters within this set were introduced in the 1983 revision, except for the range 0x2221–0x222E (kuten 2-1 to 2-14), which were part of the initial 1978 standard.
| Template:Chset-left1 | Template:Chset-cell1 | Template:Chset-cell1 | Template:Chset-cell1 | Template:Chset-cell1 | Template:Chset-cell1 | Template:Chset-cell1 | Template:Chset-cell1 | Template:Chset-cell1 | Template:Chset-cell1 | Template:Chset-cell1 | Template:Chset-cell1 | Template:Chset-cell1 | Template:Chset-cell1 | " | " | " | " | " | " | " | " | " | "
" | " | " | " | " | " | " " | " | " " | " " | " " | " " | " " | " " | " " | " " | " " | " " | " " | " " | " " | " " | " " | " " | " " | " " | " " | " " | " |
(Note: The table continues with additional symbols such as LOGICAL AND, OR, NOT SIGN, DOUBLE ARROWS, FOR ALL, THERE EXISTS, ANGLE, NABLA, APPROXIMATELY EQUAL TO, SQUARE ROOT, MUSIC SHARP/FLAT, DAGGER, PILCROW SIGN, and LARGE CIRCLE.)
Character set 0x23 (third row, digits and Roman letters)
This character set features a selection from the ISO 646 invariant repertoire, thereby also representing a portion of both ASCII and the JIS X 0201 Roman collection, excluding punctuation marks and special symbols. It includes western Arabic numerals along with uppercase and lowercase letters from the Basic Latin alphabet. When used alongside JIS X 0208 in encoding schemes such as EUC-JP, Shift JIS, or ISO 2022-JP, these characters might be mapped alternatively to the Halfwidth and Fullwidth Forms Unicode block.
For comparison, see row 3 of KPS 9566, which is identical to this row. By contrast, row 3 of KS X 1001 and row 3 of GB 2312 contain their complete national variants of ISO 646 instead of just the alphanumeric subset.
... (rest of table unchanged) }Character set 0x24 (fourth row, Hiragana)
This row consists exclusively of Japanese Hiragana syllabary characters.
Refer to row 4 of GB 2312, which corresponds precisely to this set. Additionally, rows from other standards such as row 10 in KPS 9566 and KS X 1001 employ the same arrangement but place it in a different row.
This section features Japanese Katakana symbols.
It corresponds to GB 2312’s row 5. Comparable arrangements can be found in row 11 of KPS 9566 and KS X 1001, albeit positioned in different rows. This differs from the unique Katakana configuration present in JIS X 0201.
... [table continues with same structure, omitted for brevity] ... }Code set 0x26 (row 6, Greek letters)
This row contains fundamental characters of the contemporary Greek alphabet, excluding both diacritical marks and the final sigma variant.
It matches row 6 in GB 2312 and GB 12345 as well as row 6 in KPS 9566, all sharing identical Greek symbols in this layout. It is worth noting that GB 12345 incorporates vertical text forms, whereas KPS 9566 includes Roman numerals. In contrast, KS X 1001’s row 5 shifts Greek letters to accommodate Roman numerals first.
| Reserved or unassigned code points |
|---|
Character set 0x27 (row 7, Cyrillic)
This row contains the contemporary Russian alphabet, though it might not comprehensively cover all variations of the Cyrillic script.
It aligns with row 7 in GB 2312. Comparable arrangements are found in row 12 of KS X 1001 and row 5 of KPS 9566, albeit placed differently.
... (table continues with lowercase letters and blank cells) ...
Character set 0x28 (row 8, box drawing)
All characters in this collection were introduced in 1983, absent from the initial revision of the standard in 1978.
Template:Chset-left1
Extension character set 0x2D (Row 13, NEC-specific characters)
Rows 9 through 15 within the JIS X 0208 standard remain unassigned.
Nevertheless, row 13 was designated by NEC as an extension and is widely adopted—with slight modifications—by Windows-932 (consistent with the WHATWG Encoding Standard for HTML5), the PostScript variant of MacJapanese (beginning from KanjiTalk version 7, excluding the standard version), and JIS X 0213 (the successor to JIS X 0208). Unlike other extensions for Windows-932/WHATWG and JIS X 0213, these two sets align rather than clash, thereby enhancing decoding compatibility for this particular row.
... [Table continues with units, symbols, and era names relevant for business and technical documentation] ...
Code structure
To denote code points, single-byte codes use column and line numbers, while two-byte codes employ kuten numbering. Character identifiers specify the characters independently of their code values.
Single byte codes
The majority of JIS X 0208 graphic character encodings consist of two bytes, each containing at least seven bits. In contrast, all control characters and the conventional (space) character, excluding ideographic space, utilize single-byte codes. These one-byte codes are represented by two decimal numbers: a column number and a line number, each corresponding to a hexadecimal digit. For instance, the space character's bit pattern is 010 0000 in 7-bit form or 0010 0000 in 8-bit form, which translates to column/line notation as 2/0 or hex value 0x20.
Code points and code numbers
Double-byte codes are structured into 94 distinct groups known as ku, each consisting of 94 individual codes called ten. This configuration allows for a maximum of 8,836 (94 multiplied by 94) unique code points arranged in a matrix of 94 rows by 94 columns.
A specific code point is identified by a pair of numbers—row and cell—both ranging from 1 to 94, creating a kuten coordinate expressed as "row-cell." For instance, the character "" is located at row 16, cell 1, and is represented by the code "16-01."
In the context of 7-bit JIS X 0208 (utilized within JIS X 0202 / ISO-2022-JP), both bytes span from 0x21 (corresponding to row or cell 1) up to 0x7E (row or cell 94), aligning with printable ASCII characters except for the space. When encoded, each byte value is increased by 0x20 (decimal 32); therefore, the code "16-01" for "亜" translates to 0x30 0x21. The 8-bit EUC-JP standard uses byte values between 0xA1 and 0xFE, where the high bit is set, whereas Shift JIS incorporates more intricate mappings and extends beyond row 94.
This fundamental arrangement also serves as the basis for Mainland China's GB 2312 standard, referred to as , and South Korea's KS C 5601 (currently KS X 1001), where the terms ku and ten correspond respectively to hang and yol. Subsequent standards such as JIS X 0213 introduce multiple sets of , a structure similarly adopted by CNS 11643 and related to CCCII.
Unassigned code points
Within the two-byte code space, rows numbered from 9 through 15 and from 85 through 94 are designated as unassigned code points (空き領域, aki ryōiki), indicating that no characters have been allocated in these regions. Additionally, certain cells scattered across other rows remain effectively unassigned.
These unallocated segments are generally intended to remain unused. Characters (gaiji) intended for information exchange should not be mapped to these unassigned code points unless there is explicit consensus among involved parties.
When allocating characters to these positions, it is imperative to avoid duplicating or reassigning graphic characters already defined in the standard. Special attention must be paid to issues stemming from Han unification concerning kanji glyph variations; for example, row 25 cell 66 encompasses both the prevalent "mouth" variant (高) and the less common "ladder" variant (髙) of the kanji meaning "high" or "expensive." Assigning the ladder form to an unassigned point would contravene the established standard.
In actual usage, various vendor-specific Shift JIS implementations—such as Windows-932 and MacJapanese—utilize these unallocated rows to encode proprietary extensions. Many codes left undefined in JIS X 0208 have been assigned in the more recent JIS X 0213 specification.
Character names
Every character defined in JIS X 0208 has a unique name that identifies it independently from its encoded value. These names correspond with those found in other standards like the Universal Coded Character Set (UCS/Unicode), enabling consistent character mapping. For example, both ISO/IEC 646 (US-ASCII) at column 4 line 1 and JIS X 0208 at row 3 cell 33 share the designation "LATIN CAPITAL LETTER A," signifying they represent an identical character despite differences in encoding schemes. However, certain ASCII characters—including quotation mark (2/2), apostrophe (2/7), hyphen-minus (2/13), and tilde (7/14)—are absent from this standard.
Names for non-kanji characters consist of uppercase Roman letters, spaces, and hyphens and generally follow , although some exceptions exist. These Roman-letter names function as an international naming convention similar to scientific terminology, whereas Japanese common names serve as informal references.
Kanji character names are systematically generated from their UCS/Unicode hexadecimal identifiers by prefixing "CJK UNIFIED IDEOGRAPH-". For instance, row 16 cell 1 (亜) corresponds to U+4E9C; thus, its name is "CJK UNIFIED IDEOGRAPH-4E9C." Unlike non-kanji characters, kanji do not possess Japanese common names.
Overview
The JIS X 0208 standard specifies a collection of 6,879 graphical characters encoded using two-byte sequences, referred to as the kanji set (漢字集合, kanji shūgō). This set comprises 6,355 kanji characters alongside 524 non-kanji (非漢字, hikanji) symbols, which include Latin alphabets and kana scripts.
- Special characters
- Rows 1 and 2 encompass 147 characters, featuring 18 descriptor symbols such as the ideographic space, Japanese comma, and period; eight diacritical marks including dakuten and handakuten; ten iteration marks that follow kana or kanji; 22 types of brackets; 45 mathematical operators; and 32 unit symbols like currency and postal marks.
- Numerals
- Digits from "0" through "9" are located partially within row 3.
- Latin letters
- Row 3 also contains both uppercase and lowercase English alphabets, amounting to 52 characters in total.
- Hiragana
- Row 4 consists of 83 kana characters covering unvoiced, voiced, semi-voiced sounds as well as small kana for palatalized pronunciations.
- Katakana
- Row 5 holds 86 katakana characters, including small variants of ka/ke kana and the vu kana.
- Greek letters
- Row 6 includes all 24 Greek letters in both uppercase and lowercase forms, summing to 48 characters.
- Cyrillic letters
- Row 7 features the full set of 33 Russian alphabet letters in uppercase and lowercase, totaling 66 characters.
- Box-drawing characters
- Row 8 provides 32 line elements consisting of thin, thick, and mixed styles for box drawing.
- Kanji
- Level 1 kanji occupy rows 16 through 47 with a count of 2,965 characters; level 2 kanji are found in rows 48 to 84 with a total of 3,390 characters—altogether making up the full set of 6,355 kanji.
Special characters, numerals, and Latin characters
Certain graphic characters from ISO/IEC 646:1991 IRV (ASCII) are missing in the JIS X 0208 kanji set—specifically the "QUOTATION MARK," "APOSTROPHE," "HYPHEN-MINUS," and "TILDE." The first three are represented across multiple code points within JIS X 0208 rather than as single entities, while no direct counterpart exists for the "TILDE." Instead, a "WAVE DASH" character serves as a visually similar alternative.
Subsequent standards like JIS X 0213 introduced code points for these absent ASCII symbols. Additionally, Microsoft’s Windows-31J character mapping handles these discrepancies. This division contributes to JIS X 0208 being one of the most widely adopted character sets globally but lacking upward compatibility—a significant drawback inherent to this standard.
Although approximately ninety special symbols, numerals, and Latin letters overlap between JIS X 0208 and IRV, their organization differs from ISO/IEC 646’s layout. These elements are distributed between punctuation in row 1 and letters/numbers in row 3. Only the letters and numbers in row 3 correspond directly with ISO’s arrangement.
The inclusion of full-width alphanumeric glyphs (全角英数字, zenkaku eisūji) within the kanji set highlights these incompatibilities. From its earliest versions, JIS X 0208 permitted representation of composite forms such as encircled numerals, unit ligatures for measurements, and Roman numerals via character composition rather than dedicated code points. While vendors frequently develop proprietary external characters (gaiji) to satisfy client demands, none have succeeded in officially incorporating these composites into the standard.
The revision made in the fourth edition (1997) clarified that composite characters are spacing glyphs that advance the cursor position and should not be generated through composition methods. As a result, rendering Latin letters with diacritics was generally prohibited except potentially for the ångström symbol (Å) located at row 2 cell 82.
Hiragana and katakana
Within JIS X 0208, both hiragana and katakana characters incorporate dakuten and handakuten marks as integral parts of the characters, a distinction from JIS X 0201. Additionally, this standard includes certain obsolete katakana such as and , alongside the small , which are not present in JIS X 0201.
The ordering of kana varies between these standards: JIS X 0201 arranges katakana beginning with , then lists small kana followed by full-sized kana in gojūon sequence. Conversely, JIS X 0208 organizes kana primarily by gojūon order, followed by a subdivision into "small kana, full-sized kana, kana with dakuten, and kana with handakuten," effectively grouping base kana with their modified forms. This structure supports dictionary sorting based on kana and enhances text processing capabilities in business environments.
The difference in katakana sequencing compared to JIS X 0201 is considered a factor behind the development of "half-width kana" in JIS X 0201, underscoring a limitation inherent to that earlier standard.
Kanji
The kanji repertoire in this standard is categorized into level 1 and level 2 groups, derived from four principal kanji compilations that formed the foundation of the original 6349-character set defined in the initial 1978 standard:
- – Created by the Information Processing Society of Japan kanji code committee in 1971, encompassing roughly 6086 characters.
- – Established by Japan's Administrative Management Agency in 1975, containing 2817 characters selected based on frequency and correspondence analyses for administrative data handling.
- – A now-obsolete list of 3044 characters used as a reference in the "Correspondence Analysis Results" during selection.
- – Comprises 3251 characters representing official Japanese administrative place names, compiled by the Japan Geographic Data Center.
Later updates in the second and third editions introduced six additional characters to level 2, revised character forms, and adjusted level assignments to better suit data processing requirements. These enhancements increased the standard’s effectiveness for business applications demanding extensive kanji coverage.
Level partitioning
Level 1 includes 2,965 kanji positioned within rows 16 through 47, while level 2 contains 3,390 kanji spanning rows 48 to 84.
Kanji in level 1 were chosen based on their prevalence across multiple glyph sources, referencing lists such as the tōyō kanji, its correction draft, and the jinmeiyō kanji. Standards like JIS C 6260 (prefecture codes) and JIS C 6261 (municipal codes) were consulted to ensure inclusion of nearly all kanji used for Japanese administrative divisions within level 1. Expert review further refined this selection.
Level 2 encompasses kanji found in these references but excluded from level 1 due to factors like ambiguous readings or lower usage frequency at selection time. Typically, level 1 contains more commonly used kanji whereas level 2 covers those less frequent; nonetheless, some level 2 kanji such as "to soar" (翔) and "to glitter" (煌) have become more prevalent over time. Conversely, certain level 1 kanji like those denoting "centimeter" (糎) and "millimeter" (粍) have decreased in usage. Among the current jōyō kanji, thirty are classified under level 2 with three not included at all; regarding jinmeiyō kanji, 192 appear in level 2 while another 105 fall outside this standard.
Arrangement
Level 1 kanji are organized according to a specified "representative reading," which can be either an on or kun pronunciation, arranged in the order of the gojūon sequence. Typically, the main on reading is selected; if this is missing or rarely used, the kun reading is chosen instead, with verbs represented in their ren'yōkei (continuative) form.
For example, the cells from 1 to 41 in row 16 contain readings that start with the vowel a. Within this group, some characters adopt their kun readings as representatives, such as 葵 (which has the on reading "ki" and kun reading "aoi") and 粟 (with on readings "zoku" or "shoku" and kun reading "awa"). Verbs like 逢 (with on "hō" and kun "a(i)") and 扱 (with on readings "sō" or "kyū" and kun "atsuka(i)") utilize their ren'yōkei forms.
When multiple kanji share an identical representative reading, those with an on pronunciation are listed before those with a kun reading. If readings coincide further, the sorting then depends on the primary radical followed by stroke count.
In both levels, itaiji are positioned immediately following their standard counterparts. For instance, in level 2, after 劍 (located at row 49 cell 88), three variants (劔, 劒, 剱) appear consecutively despite stroke count differences. The principal variant 剣 is found in level 1 (row 23 cell 85), while another variant (釼) is categorized under the gold radical in level 2.
Kanji in level 2 are mainly ordered according to their radical and number of strokes; if ties occur, reading order serves as a tiebreaker.
Kanji from unknown sources
| Kuten | Symbol | Classification |
|---|---|---|
| 52-55 | 墸 | Unknown |
| 52-63 | 壥 | Unknown |
| 54-12 | 妛 | Source unclear |
| 55-27 | 彁 | Unidentifiable |
| 57-43 | 挧 | Source unclear |
| 58-83 | 暃 | Source unclear |
| 59-91 | 椦 | Source unclear |
| 60-57 | 槞 | Source unclear |
| 74-12 | 蟐 | Source unclear |
| 74-57 | 袮 | Source unclear |
| 79-64 | 閠 | Source unclear |
| 81-50 | 駲 | Source unclear |
A number of kanji incorporated into this set do not appear in comprehensive kanji dictionaries and have origins that remain unidentified. For example, Tajima (1979) documented sixty-three such characters absent from major references like Shinjigen or Dai Kan-Wa jiten, which also cannot be explained as ryakuji. These characters became known as or .
The committee responsible for drafting the fourth standard examined these anomalies and found that the original committee heavily depended on the "Correspondence Analysis Results" when compiling kanji. Many of these untraceable characters were sourced from lists such as "Japanese Personality Registration Name Kanji" and "Kanji for National Administrative District Listing," which were part of those results.
No original documentation exists for the "Japanese Personality Registration Name Kanji." Regarding the "National Administrative District Listing," extensive investigation—including reviewing historical documents and analyzing NTT phone book name data—was carried out. This effort reduced the number of unexplained kanji to twelve, as shown above. Some glyphs likely originated from transcription errors; for instance, 妛 probably resulted from a miscopying of 𡚴 by combining 山 and 女 components, with a printing flaw misinterpreted as an additional stroke (as explained in the Jōyō kanji jiten).
Consolidation of Kanji Variants
The fourth edition of the standard (1997) introduces the concept of unification (包摂, hōsetsu), which assigns a single code point to a character irrespective of its variant forms. This standard restricts permissible glyphs and explicitly outlines how specific allographic glyphs are merged into a unified graphemic code point.
A represents the abstract graphical identity of a character, while a refers to the tangible shape a glyph assumes, such as in handwritten or printed forms. Differences among character manifestations within one glyph are termed .
Unification hinges on the code point's and the applied. The exemplar glyph establishes the baseline, with any glyphs featuring parts substituted according to criteria also corresponding to that same code point.
For example, the exemplar glyph at 33-46 (僧) consists of radical 9 (亻) combined with the kanji 曽, from which the kana so is derived. Unification criterion 101 identifies three variants: the commonly used Japanese 曽, the traditional 曾 incorporating radical 12 (八), and an inverted form of radical 12 (曾). All three variants share the identical code point at 33-46.
The fourth standard specifies 186 unification criteria. When an exemplar glyph includes multiple components, each may be subject to one criterion; however, overlapping criteria that replicate another code point's glyph are prohibited.
Exemplar glyphs function as reference points for code points but do not represent officially endorsed forms. The unification criteria primarily serve to guide the assignment of frequently used kanji; creating rarely used kanji based on these guidelines is discouraged.
Kanji selection within the set does not consistently align with unification criteria. For instance, 41-7 encompasses both crossed and uncrossed stroke versions of 彥/彦 according to criterion 72, whereas 20-73 corresponds solely to the uncrossed 顔 form, and 80-90 corresponds exclusively to the crossed 顏 form.
The terminology "unification," "unification criteria," and "exemplar glyph" was introduced in this fourth standard. Earlier editions categorized kanji relationships as "independent" (独立), "compatible" (対応), and "equivalent" (同値), merging equivalent characters—including those differing slightly stylistically—into a single code point.
Since the initial standard, it has been declared that specific character form designs fall outside the scope; emphasis is placed on character concepts and coding rather than exact visual representation. This principle persists through all subsequent revisions, including JIS X 0208:1997.
Unification Criteria for Compatibility
The fourth edition defines unification criteria aimed at preserving compatibility with earlier standards (過去の規格との互換性を維持するための包摂規準). These apply to 29 code points exhibiting notable glyph differences between JIS C 6226-1983 and JIS C 6226-1978. For these cases, glyphs from the 1983 standard are designated as "A," while those from 1978 are labeled "B." Both forms may be utilized; however, when asserting compatibility, it is necessary to specify which variant is applied for each code point.
Encoding Methods Defined by JIS X 0208
Article 7 and appendices 1 and 2 of JIS X 0208:1997 delineate eight distinct encoding schemes.
The code areas include: "CL" (control left) spanning 0/0–1/15, "GL" (graphic left) covering 2/1–7/14, "CR" (control right) from 8/0–9/15, and "GR" (graphic right) ranging from 10/1–15/14. Code 2/0 represents the graphic character "SPACE," while 7/15 corresponds to the control character "DELETE." The C0 control characters from JIS X 0211 (aligned with ISO/IEC 6429) occupy the CL region.
- 7-bit kanji encoding
- Specified within the standard; assigns JIS X 0208 double-byte set to GL.
- 8-bit kanji encoding
- Similar to the 7-bit version but employs eight-bit bytes; CR may be unused or encode C1 control characters from JIS X 0211; GR remains unused.
- International Reference Version plus 7-bit kanji encoding
- IRV (ISO/IEC 646:1991, equivalent to US-ASCII) allocated to GL via shift-in control; shift-out switches to JIS X 0208 double-byte set.
- Latin characters plus 7-bit kanji encoding
- Comparable to IRV+7-bit but substitutes IRV with ISO/IEC 646:JP (JISCII Roman set).
- International Reference Version plus 8-bit kanji encoding
- IRV assigned to GL and JIS X 0208 set assigned to GR; effectively forms a subset of EUC-JP excluding half-width katakana and supplementary kanji.
- Latin characters plus 8-bit kanji encoding
- Mirrors IRV+8-bit but replaces IRV with ISO/IEC 646:JP.
- Shift-coded character set
- Outlined in Appendix 1 as Shift-Coded Representation (シフト符号化表現), this serves as the definitive specification of Shift JIS.
- RFC 1468-coded character set
- Presented in Appendix 2 resembling ISO-2022-JP but utilizing eight-bit bytes instead of seven.
Among these options, only the Shift-coded character set holds registration by the IANA. The remaining encodings closely relate to IANA-registered formats such as EUC-JP and ISO-2022-JP.
Escape sequences for JIS X 0202 / ISO 2022
Within ISO 2022/JIS X 0202 (such as ISO-2022-JP), JIS X 0208 is utilized. The escape sequences that specify JIS X 0208 for each of the four ISO 2022 code sets are detailed below, where "ESC" denotes the control character "Escape" (0x1B).
| Standard | G0 | G1 | G2 | G3 |
|---|---|---|---|---|
| 78 | ESC 2/4 4/0 | ESC 2/4 2/9 4/0 | ESC 2/4 2/10 4/0 | ESC 2/4 2/11 4/0 |
| 83 | ESC 2/4 4/2 | ESC 2/4 2/9 4/2 | ESC 2/4 2/10 4/2 | ESC 2/4 2/11 4/2 |
| 90 onward | ESC 2/6 4/0 ESC 2/4 4/2 | ESC 2/6 4/0 ESC 2/4 2/9 4/2 | ESC 2/6 4/0 ESC 2/4 2/10 4/2 | ESC 2/6 4/0 ESC 2/4 2/11 4/2 |
The sequence ESC 2/4 indicates selection of a multi-byte character set, while ESC 2/6 designates a revision. The identifier byte for JIS C 6226:1978 is 4/0 (@), whereas JIS C 6226:1983 / JIS X 0208:1983 and JIS X 0208:1990 use byte 4/2 (B), with the latter distinguished by the revision marker 4/0.
Duplicate encodings of ASCII and JIS X 0201
When the kanji character set is combined with ISO/IEC 646:1991 IRV (ASCII) or the Latin characters from JIS X 0201 (JIS-Roman), overlapping characters result in duplicated encodings due to different code point assignments.
JIS X 0208:1997 generally prohibits using duplicate kanji set code points for commonly shared characters, treating identically named characters as equivalent. For example, "LATIN CAPITAL LETTER A" exists both in ASCII (bit pattern 4/1) and in the kanji set (row 3 cell 33); the standard forbids employing the kanji set's code point to prevent duplication.
Nonetheless, for compatibility with older systems, kanji set code points can be interpreted as full-width variants distinct from ASCII or JIS-Roman. For instance, bit pattern 10/3 12/1 may represent a full-width "A".
Even strict adherence to the standard does not guarantee unique encoding. The hyphen, for example, can be encoded either as ASCII bit pattern 2/13 ("HYPHEN-MINUS") or as kanji set row 1 cell 30 ("HYPHEN"), with no clear preference specified. Similar ambiguities exist for minus signs and quotation marks.
Moreover, the presence of both full-width "IDEOGRAPHIC SPACE" (row 1 cell 1) and half-width space (bit pattern 2/0) exemplifies unresolved distinctions left unspecified by the standard.
History
Japanese Industrial Standards are subject to reaffirmation, revision, or withdrawal within five years after their initial release. Since its introduction, this standard has undergone three revisions; currently, the fourth edition remains in effect.
First standard
The original standard, JIS C 6226-1978, was promulgated by the Japanese Minister of International Trade and Industry on January 1, 1978, also referred to as 78JIS. It was developed by a committee from JIPDEC led by Moriguchi Shigeichi. The set comprised a total of 6,802 characters: including 453 non-Kanji elements (such as Hiragana, Katakana, Roman, Greek, Cyrillic alphabets, and punctuation) alongside 6,349 Kanji characters (divided into level 1 with 2,965 characters and level 2 with 3,384). At this stage, box-drawing characters were not incorporated. The typeface employed was Ishii Mincho from Shaken Co., Ltd.
Second standard
The second edition, JIS C 6226-1983, often referred to as 83JIS, was released on 1 September 1983. This version was developed by a JIPDEC kanji code committee under the leadership of Motooka Tōru. It incorporated updates reflecting the official adoption of jōyō kanji, the enforcement of jinmeiyō kanji, and harmonization with Japanese Teletex protocols.
Notable modifications included:
- Inclusion of special characters
- Based on recommendations from JICST and standards such as JIS Z 8201-1981 (mathematical symbols) and JIS Z 8202-1982 (units and chemical symbols), 39 special characters were added.
- Introduction of box-drawing characters
- A set of 32 new box-drawing characters was incorporated.
- Reassignment of itaiji code points
- To enhance consistency, code points for 22 variant kanji pairs were exchanged between levels 1 and 2.
- Adjustments in level 2 kanji
- Four characters were relocated to new code points in row 84 to accommodate alternate forms.
- Refinement of character shapes
- Approximately 300 kanji glyphs underwent revisions, predominantly favoring simplified variants over Kangxi-style forms within level 1 kanji. Examples include changes at row 18 cell 10 (鷗 replaced by 鴎) and row 38 cell 34 (瀆 replaced by 涜).
Simplifications primarily impacted level 1 kanji more than those in level 2. In certain cases, glyphs shifted towards or away from Kangxi-style designs to better represent their original meanings. The differences in glyph forms between levels prompted the formulation of unification criteria in the fourth standard. For 29 code points where modifications conflicted with the initial standard’s intent, specific unification rules were established to maintain backward compatibility.
With the emergence of the "X" series classification for information-related standards, this edition was redesignated as JIS X 0208-1983 on 1 March 1987.
Third standard
The third edition, JIS X 0208-1990, superseded the second on 1 September 1990 and is also known as 90JIS. The draft was prepared by a committee from the Japanese Standards Association, chaired by Tajima Kazuo, under AIST's commission.
This update involved modifications to 225 kanji glyphs and added two new characters to level 2 (positions 84-05 "" and 84-06 ""). Additionally, it disunified certain itaiji variants previously grouped together (such as at positions 49-59 "" and 63-70 ""). Some adjustments corresponded with the inclusion of the 118 jinmeiyō kanji introduced in March 1990. The standard utilized Heisei Mincho font for its character set.
Fourth standard
Published on 20 January 1997, the fourth revision JIS X 0208:1997—also referred to as 97JIS—updated the third edition. This version was drafted by a JSA committee led by chairman Shibano Kōji under AIST's direction.
Rather than altering the character repertoire or example glyphs, this revision concentrated on clarifying provisions and enhancing practical usability. While the third standard comprised only 65 pages excluding explanatory notes, the fourth expanded significantly to 374 pages due to extensive rewrites and supplementary content.
Key enhancements included:
- Specification of encoding techniques
- Previously limited to JIS X 0202-based encoding using escape sequences, the fourth edition introduced encoding methods that operate without escape sequences for extending codes.
- Clarification regarding unassigned code points
- The third edition ambiguously permitted gaiji assignments to certain unallocated code points. The fourth explicitly forbade general use of such unassigned codes while setting conditions for exceptional cases.
- Removal of redundant encodings
- Each character was assigned a unique "character name" linked to other standards. When combined with ISO/IEC 646 or JIS X 0201, only one code point per character name is allowed, eliminating duplicate encodings.
- Examination of kanji origins
- Characters absent from both the Kangxi Dictionary and Dai Kanwa Jiten were identified, prompting reconsideration of their inclusion rationale from the first standard.
- Establishment of kanji unification guidelines
- The original intent concerning glyph scope per code point was reaffirmed based on draft materials, with clear criteria defined for unifying kanji glyphs.
- Incorporation of prevalent encoding standards
- By this time, Shift JIS and ISO-2022-JP had become widely adopted for personal computers and email communication. These were formally integrated as "Shift-Coded Representation" and "RFC 1468-Coded Representation," respectively.
Successors
JIS X 0213 (extended kanji) was developed to offer a more comprehensive character collection for encoding contemporary Japanese, building upon the kanji repertoire established by JIS X 0208. The standard's creators recommend transitioning from JIS X 0208 to JIS X 0213, emphasizing its alignment with the Hyōgai Kanji Glyph List and the updated jinmeiyō kanji.
Despite being introduced in 2000, its uptake has been gradual. The drafting committee in 2004 observed that "most information systems continue to rely solely on JIS X 0208." Regarding Microsoft Windows, the leading operating system in personal computing, support for JIS X 0213 was introduced starting with Windows Vista (2006). Meanwhile, Mac OS X has supported it since version 10.1 (2001), and numerous Unix-like platforms such as Linux provide optional support. This indicates that personal computer compatibility is unlikely to be a barrier to eventual adoption.
Several contributors anticipate that both JIS X 0208 and JIS X 0213 will coexist for an extended period. The following challenges inhibit JIS X 0213 from fully replacing JIS X 0208:
- Character sets used in Japanese mobile phone systems remain based on JIS X 0208, with no current plans for migration, restricting adoption within a crucial communication channel.
- JIS X 0213 does not maintain complete upward compatibility with the unification principles of JIS X 0208, complicating conversions for extensive archives (such as bibliographic databases and Aozora Bunko) that depend on strict standardization.
- Numerous systems utilize unassigned code points within JIS X 0208—for instance, Windows incorporates IBM and NEC extended characters, while mobile devices allocate code points for emoji. These assignments conflict with those in JIS X 0213, making transitions more complex. Although there are strategies to migrate via the UCS/Unicode, hesitation persists until implementations stabilize.
- The improvements offered by JIS X 0213 primarily address less commonly used characters and require nearly twice as many glyphs for marginal practical advantage—a low return on investment for organizations with limited resources.
Implementations
Because JIS X 0208 / JIS C 6226 specifies a character repertoire rather than a fixed encoding method, various companies have created proprietary encodings:
- Apple: MacJapanese (based on Shift_JIS)
- Fujitsu: JEF kanji code (EBCDIC-based)
- Hitachi: KEIS (EBCDIC-based)
- IBM: IBM-932 and IBM-942 (Shift_JIS-based)
- Microsoft: Windows-932 (Shift_JIS-based)
- NEC: JIPS
Certain encodings such as Windows-932, MacJapanese, and NEC's PC98 format incorporate vendor-specific characters within code points unallocated by the standard. IBM-932 and IBM-942 also include vendor-defined characters but outside of the designated JIS X 0208 area.
ISO/IEC 646 IRV and ASCII
The kanji collection is not fully compatible upwards with ISO/IEC 646:1991 IRV (ASCII), yet they can be combined according to specifications in JIS X 0208 (utilizing IRV alongside either 7-bit or 8-bit kanji codes) as well as in EUC-JP.
JIS X 0201
While JIS X 0208 lacks three Latin letters present in JIS X 0201, it contains all katakana characters defined by JIS X 0201. The kanji set can be integrated with Latin letters following the guidelines of JIS X 0208, including through shift-coded character sets like Shift JIS. Additionally, kanji and katakana sets may be combined within EUC-JP encoding.
JIS X 0212
JIS X 0212 specifies an auxiliary kanji collection designed to meet information processing requirements beyond those covered by JIS X 0208, establishing a separate 94-by-94 kanji grid instead of enlarging the primary set. This standard can be utilized in conjunction with JIS X 0208 within the EUC-JP encoding framework, and both serve as foundational references for Unicode’s Han unification process.
Compared to the second revision of JIS X 0208, JIS X 0212 includes 28 code points representing pre-modification character forms. Additionally, it reclassifies the "closure mark" from a non-kanji character in JIS X 0208 to a kanji. Other than this reassignment, there is no overlap in characters with JIS X 0208, which restricts its independent applicability.
The fourth edition of JIS X 0208 deliberately omits any formal linkage to JIS X 0212, reflecting the drafting committee’s critical stance on JIS X 0212’s character selection methodology and documentation quality. The committee identified substantial issues with how characters were chosen and integrated, ultimately deeming combined usage impractical and leaving any relationship undefined.
JIS X 0213
JIS X 0213 broadens the kanji repertoire established by JIS X 0208 to more comprehensively represent contemporary Japanese text. It defines a total of 11,233 characters—comprising 1,183 non-kanji and 10,050 kanji—distributed over two distinct 94-by-94 planes. The first plane extends the non-kanji characters and levels 1 through 3 kanji from JIS X 0208, while the second plane incorporates level 4 kanji, occupying previously unused rows within JIS X 0212’s layout to maintain compatibility with EUC-JP.
JIS X 0213 also introduces Shift_JISx0213, a variant of Shift_JIS encoding that encompasses its entire character set.
Although the first plane of JIS X 0213 largely replaces JIS X 0208, it applies alternative unification rules for specific kanji. Certain glyphs that were unified under a single code point in JIS X 0208 are allocated distinct code points in JIS X 0213. For instance, the kanji located at row 33 cell 46 in JIS X 0208 is divided into two separate forms in JIS X 0213 plane 1 (row 33 cell 46 and row 14 cell 41). This differentiation introduces ambiguity that complicates automated mapping between these standards and impacts backward compatibility—a concern noted by the drafting committee.
Despite this, the majority of code points align directly between JIS X 0208 and plane 1 of JIS X 0213, reducing practical confusion since font designs usually adhere to JIS X 0208 glyph models and end-users are typically unaware of these unification distinctions.
ISO/IEC 10646 and Unicode
The kanji collection defined by JIS X 0208 serves as a key reference for the Han unification effort within ISO/IEC 10646 (Universal Character Set) and Unicode. Each kanji from JIS X 0208 corresponds uniquely to a code point within the Basic Multilingual Plane (BMP) of UCS/Unicode.
Non-kanji symbols from JIS X 0208 also map to BMP code points; however, certain implementations assign mappings that differ from those in UCS/Unicode, which are based on character names specified in the JIS X 0208:1997 standard.
Reference footnotes
See also
- JIS coded character sets
- JIS X 0201 "7-bit and 8-bit coded character sets for information interchange"
- JIS X 0202 "Information technology – Character code structure and extension techniques" (ISO/IEC 2022)
- JIS X 0208 "7-bit and 8-bit double byte coded KANJI sets for information interchange"
- JIS X 0211 "Control functions for coded character sets" (ISO/IEC 6429)
- JIS X 0212 "Code of the supplementary Japanese graphic character set for information interchange"
- JIS X 0213 "7-bit and 8-bit double byte coded extended KANJI sets for information interchange"
- JIS X 0221 "Universal Multiple-Octet Coded Character Set (UCS)" (ISO/IEC 10646)
- Extended shinjitai
- Help:Japanese
References
When citing, Japanese personal names are shown in Western sequence if Romanized, but retain the traditional Eastern order otherwise.
- Nishimura, Hirohiko [], 1978. The Kanji JIS []. Standardization Journal [], 171: 3–8.
- Nomura, Masaaki [], 1984. Amendments to JIS C 6226: Kanji codes for data interchange []. Standardization Journal [], 14 (3): 4–9.
- Ogata, Katsuhiro [], 2006a. Glyphs not unified in 97JIS among those modified in JIS C 6226-1983 (83JIS) [] (retrieved 29 January 2007).
- Ogata, Katsuhiro [], 2006b. Glyphs within the unification range altered in JIS C 6226-1983 (83JIS) [] (retrieved 29 January 2007).
- Satō, Takayuki [], 2004. Updates to JIS X 0213 (expanded Kanji sets) []. Standardization Journal [], 34 (4): 8–12.
- Shibano, Kōji [], various years. Principal publications concerning revisions and expansions of JIS kanji standards.
- Tajima, Kazuo [], 1979. Design considerations and management of kanji in processing systems. Journal of Information Processing Society of Japan, 21 (10):753–761.
- Uchida, Tomio [],1990. Introduction of JIS X 0212 supplementary kanji codes. Standardization Journal, 20 (11):6–11.
- Yasuoka, Kōichi [],2001a,b & Motoko Yasuoka,2006. Research on recent character coding developments in Japan and historical distinctions in JIS kanji plans.
External links
- The International Register maintained by IPSJ/ITSCJ.
- Japanese Industrial Standards Committee database (access current standards).
- Japanese Standards Association database (to acquire latest standards).
- Unification rules within JIS X 0208 and X 0213 standards
- Cyber Librarian – comprehensive JIS kanji index