7 Network Working Group A. Phillips, Ed.
8 Request for Comments: 4646 Yahoo! Inc.
10 Obsoletes: 3066 Google
11 Category: Best Current Practice September 2006
14 Tags for Identifying Languages
18 This document specifies an Internet Best Current Practices for the
19 Internet Community, and requests discussion and suggestions for
20 improvements. Distribution of this memo is unlimited.
24 Copyright (C) The Internet Society (2005).
28 This document describes the structure, content, construction, and
29 semantics of language tags for use in cases where it is desirable to
30 indicate the language used in an information object. It also
31 describes how to register values for use in language tags and the
32 creation of user-defined extensions for private interchange. This
33 document, in combination with RFC 4647, replaces RFC 3066, which
58 Phillips & Davis Best Current Practice [Page 1]
60 RFC 4646 Tags for Identifying Languages September 2006
65 1. Introduction ....................................................3
66 2. The Language Tag ................................................4
67 2.1. Syntax .....................................................4
68 2.2. Language Subtag Sources and Interpretation .................7
69 2.2.1. Primary Language Subtag .............................8
70 2.2.2. Extended Language Subtags ..........................10
71 2.2.3. Script Subtag ......................................11
72 2.2.4. Region Subtag ......................................11
73 2.2.5. Variant Subtags ....................................13
74 2.2.6. Extension Subtags ..................................14
75 2.2.7. Private Use Subtags ................................16
76 2.2.8. Preexisting RFC 3066 Registrations .................16
77 2.2.9. Classes of Conformance .............................17
78 3. Registry Format and Maintenance ................................18
79 3.1. Format of the IANA Language Subtag Registry ...............18
80 3.2. Language Subtag Reviewer ..................................24
81 3.3. Maintenance of the Registry ...............................24
82 3.4. Stability of IANA Registry Entries ........................25
83 3.5. Registration Procedure for Subtags ........................29
84 3.6. Possibilities for Registration ............................32
85 3.7. Extensions and Extensions Registry ........................34
86 3.8. Initialization of the Registries ..........................37
87 4. Formation and Processing of Language Tags ......................38
88 4.1. Choice of Language Tag ....................................38
89 4.2. Meaning of the Language Tag ...............................40
90 4.3. Length Considerations .....................................41
91 4.3.1. Working with Limited Buffer Sizes ..................42
92 4.3.2. Truncation of Language Tags ........................43
93 4.4. Canonicalization of Language Tags .........................44
94 4.5. Considerations for Private Use Subtags ....................45
95 5. IANA Considerations ............................................46
96 5.1. Language Subtag Registry ..................................46
97 5.2. Extensions Registry .......................................47
98 6. Security Considerations ........................................48
99 7. Character Set Considerations ...................................48
100 8. Changes from RFC 3066 ..........................................49
101 9. References .....................................................52
102 9.1. Normative References ......................................52
103 9.2. Informative References ....................................53
104 Appendix A. Acknowledgements ......................................55
105 Appendix B. Examples of Language Tags (Informative) ...............56
114 Phillips & Davis Best Current Practice [Page 2]
116 RFC 4646 Tags for Identifying Languages September 2006
121 Human beings on our planet have, past and present, used a number of
122 languages. There are many reasons why one would want to identify the
123 language used when presenting or requesting information.
125 A user's language preferences often need to be identified so that
126 appropriate processing can be applied. For example, the user's
127 language preferences in a Web browser can be used to select Web pages
128 appropriately. Language preferences can also be used to select among
129 tools (such as dictionaries) to assist in the processing or
130 understanding of content in different languages.
132 In addition, knowledge about the particular language used by some
133 piece of information content might be useful or even required by some
134 types of processing; for example, spell-checking, computer-
135 synthesized speech, Braille transcription, or high-quality print
138 One means of indicating the language used is by labeling the
139 information content with an identifier or "tag". These tags can be
140 used to specify user preferences when selecting information content,
141 or for labeling additional attributes of content and associated
144 Tags can also be used to indicate additional language attributes of
145 content. For example, indicating specific information about the
146 dialect, writing system, or orthography used in a document or
147 resource may enable the user to obtain information in a form that
148 they can understand, or it can be important in processing or
149 rendering the given content into an appropriate form or style.
151 This document specifies a particular identifier mechanism (the
152 language tag) and a registration function for values to be used to
153 form tags. It also defines a mechanism for private use values and
156 This document, in combination with [RFC4647], replaces [RFC3066],
157 which replaced [RFC1766]. For a list of changes in this document,
160 The keywords "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
161 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
162 document are to be interpreted as described in [RFC2119].
170 Phillips & Davis Best Current Practice [Page 3]
172 RFC 4646 Tags for Identifying Languages September 2006
177 Language tags are used to help identify languages, whether spoken,
178 written, signed, or otherwise signaled, for the purpose of
179 communication. This includes constructed and artificial languages,
180 but excludes languages not intended primarily for human
181 communication, such as programming languages.
185 The language tag is composed of one or more parts, known as
186 "subtags". Each subtag consists of a sequence of alphanumeric
187 characters. Subtags are distinguished and separated from one another
188 by a hyphen ("-", ABNF [RFC4234] %x2D). A language tag consists of a
189 "primary language" subtag and a (possibly empty) series of subsequent
190 subtags, each of which refines or narrows the range of languages
191 identified by the overall tag.
193 Usually, each type of subtag is distinguished by length, position in
194 the tag, and content: subtags can be recognized solely by these
195 features. The only exception to this is a fixed list of
196 grandfathered tags registered under RFC 3066 [RFC3066]. This makes
197 it possible to construct a parser that can extract and assign some
198 semantic information to the subtags, even if the specific subtag
199 values are not recognized. Thus, a parser need not have an up-to-
200 date copy (or any copy at all) of the subtag registry to perform most
201 searching and matching operations.
226 Phillips & Davis Best Current Practice [Page 4]
228 RFC 4646 Tags for Identifying Languages September 2006
231 The syntax of the language tag in ABNF [RFC4234] is:
233 Language-Tag = langtag
234 / privateuse ; private use tag
235 / grandfathered ; grandfathered registrations
244 language = (2*3ALPHA [ extlang ]) ; shortest ISO 639 code
245 / 4ALPHA ; reserved for future use
246 / 5*8ALPHA ; registered language subtag
248 extlang = *3("-" 3ALPHA) ; reserved for future use
250 script = 4ALPHA ; ISO 15924 code
252 region = 2ALPHA ; ISO 3166 code
253 / 3DIGIT ; UN M.49 code
255 variant = 5*8alphanum ; registered variants
258 extension = singleton 1*("-" (2*8alphanum))
260 singleton = %x41-57 / %x59-5A / %x61-77 / %x79-7A / DIGIT
261 ; "a"-"w" / "y"-"z" / "A"-"W" / "Y"-"Z" / "0"-"9"
262 ; Single letters: x/X is reserved for private use
264 privateuse = ("x"/"X") 1*("-" (1*8alphanum))
266 grandfathered = 1*3ALPHA 1*2("-" (2*8alphanum))
267 ; grandfathered registration
268 ; Note: i is the only singleton
269 ; that starts a grandfathered tag
271 alphanum = (ALPHA / DIGIT) ; letters and numbers
273 Figure 1: Language Tag ABNF
275 Note: There is a subtlety in the ABNF for 'variant': variants
276 starting with a digit MAY be four characters long, while those
277 starting with a letter MUST be at least five characters long.
282 Phillips & Davis Best Current Practice [Page 5]
284 RFC 4646 Tags for Identifying Languages September 2006
287 All subtags have a maximum length of eight characters and whitespace
288 is not permitted in a language tag. For examples of language tags,
291 Note that although [RFC4234] refers to octets, the language tags
292 described in this document are sequences of characters from the
293 US-ASCII [ISO646] repertoire. Language tags MAY be used in documents
294 and applications that use other encodings, so long as these encompass
295 the US-ASCII repertoire. An example of this would be an XML document
296 that uses the UTF-16LE [RFC2781] encoding of [Unicode].
298 The tags and their subtags, including private use and extensions, are
299 to be treated as case insensitive: there exist conventions for the
300 capitalization of some of the subtags, but these MUST NOT be taken to
305 o [ISO639-1] recommends that language codes be written in lowercase
308 o [ISO3166-1] recommends that country codes be capitalized ('MN'
311 o [ISO15924] recommends that script codes use lowercase with the
312 initial letter capitalized ('Cyrl' Cyrillic).
314 However, in the tags defined by this document, the uppercase US-ASCII
315 letters in the range 'A' through 'Z' are considered equivalent and
316 mapped directly to their US-ASCII lowercase equivalents in the range
317 'a' through 'z'. Thus, the tag "mn-Cyrl-MN" is not distinct from
318 "MN-cYRL-mn" or "mN-cYrL-Mn" (or any other combination), and each of
319 these variations conveys the same meaning: Mongolian written in the
320 Cyrillic script as used in Mongolia.
322 Although case distinctions do not carry meaning in language tags,
323 consistent formatting and presentation of the tags will aid users.
324 The format of the tags and subtags in the registry is RECOMMENDED.
325 In this format, all non-initial two-letter subtags are uppercase, all
326 non-initial four-letter subtags are titlecase, and all other subtags
338 Phillips & Davis Best Current Practice [Page 6]
340 RFC 4646 Tags for Identifying Languages September 2006
343 2.2. Language Subtag Sources and Interpretation
345 The namespace of language tags and their subtags is administered by
346 the Internet Assigned Numbers Authority (IANA) [RFC2860] according to
347 the rules in Section 5 of this document. The Language Subtag
348 Registry maintained by IANA is the source for valid subtags: other
349 standards referenced in this section provide the source material for
352 Terminology in this section:
354 o Tag or tags refers to a complete language tag, such as
355 "fr-Latn-CA". Examples of tags in this document are enclosed in
356 double-quotes ("en-US").
358 o Subtag refers to a specific section of a tag, delimited by hyphen,
359 such as the subtag 'Latn' in "fr-Latn-CA". Examples of subtags in
360 this document are enclosed in single quotes ('Latn').
362 o Code or codes refers to values defined in external standards (and
363 that are used as subtags in this document). For example, 'Latn'
364 is an [ISO15924] script code that was used to define the 'Latn'
365 script subtag for use in a language tag. Examples of codes in
366 this document are enclosed in single quotes ('en', 'Latn').
368 The definitions in this section apply to the various subtags within
369 the language tags defined by this document, excepting those
370 "grandfathered" tags defined in Section 2.2.8.
372 Language tags are designed so that each subtag type has unique length
373 and content restrictions. These make identification of the subtag's
374 type possible, even if the content of the subtag itself is
375 unrecognized. This allows tags to be parsed and processed without
376 reference to the latest version of the underlying standards or the
377 IANA registry and makes the associated exception handling when
378 parsing tags simpler.
380 Subtags in the IANA registry that do not come from an underlying
381 standard can only appear in specific positions in a tag.
382 Specifically, they can only occur as primary language subtags or as
385 Note that sequences of private use and extension subtags MUST occur
386 at the end of the sequence of subtags and MUST NOT be interspersed
387 with subtags defined elsewhere in this document.
389 Single-letter and single-digit subtags are reserved for current or
390 future use. These include the following current uses:
394 Phillips & Davis Best Current Practice [Page 7]
396 RFC 4646 Tags for Identifying Languages September 2006
399 o The single-letter subtag 'x' is reserved to introduce a sequence
400 of private use subtags. The interpretation of any private use
401 subtags is defined solely by private agreement and is not defined
402 by the rules in this section or in any standard or registry
403 defined in this document.
405 o All other single-letter subtags are reserved to introduce
406 standardized extension subtag sequences as described in
409 The single-letter subtag 'i' is used by some grandfathered tags, such
410 as "i-enochian", where it always appears in the first position and
411 cannot be confused with an extension.
413 2.2.1. Primary Language Subtag
415 The primary language subtag is the first subtag in a language tag
416 (with the exception of private use and certain grandfathered tags)
417 and cannot be omitted. The following rules apply to the primary
420 1. All two-character language subtags were defined in the IANA
421 registry according to the assignments found in the standard ISO
422 639 Part 1, "ISO 639-1:2002, Codes for the representation of
423 names of languages -- Part 1: Alpha-2 code" [ISO639-1], or using
424 assignments subsequently made by the ISO 639 Part 1 maintenance
425 agency or governing standardization bodies.
427 2. All three-character language subtags were defined in the IANA
428 registry according to the assignments found in ISO 639 Part 2,
429 "ISO 639-2:1998 - Codes for the representation of names of
430 languages -- Part 2: Alpha-3 code - edition 1" [ISO639-2], or
431 assignments subsequently made by the ISO 639 Part 2 maintenance
432 agency or governing standardization bodies.
434 3. The subtags in the range 'qaa' through 'qtz' are reserved for
435 private use in language tags. These subtags correspond to codes
436 reserved by ISO 639-2 for private use. These codes MAY be used
437 for non-registered primary language subtags (instead of using
438 private use subtags following 'x-'). Please refer to Section 4.5
439 for more information on private use subtags.
441 4. All four-character language subtags are reserved for possible
442 future standardization.
444 5. All language subtags of 5 to 8 characters in length in the IANA
445 registry were defined via the registration process in Section 3.5
446 and MAY be used to form the primary language subtag. At the time
450 Phillips & Davis Best Current Practice [Page 8]
452 RFC 4646 Tags for Identifying Languages September 2006
455 this document was created, there were no examples of this kind of
456 subtag and future registrations of this type will be discouraged:
457 primary languages are strongly RECOMMENDED for registration with
458 ISO 639, and proposals rejected by ISO 639/RA will be closely
459 scrutinized before they are registered with IANA.
461 6. The single-character subtag 'x' as the primary subtag indicates
462 that the language tag consists solely of subtags whose meaning is
463 defined by private agreement. For example, in the tag "x-fr-CH",
464 the subtags 'fr' and 'CH' SHOULD NOT be taken to represent the
465 French language or the country of Switzerland (or any other value
466 in the IANA registry) unless there is a private agreement in
467 place to do so. See Section 4.5.
469 7. The single-character subtag 'i' is used by some grandfathered
470 tags (see Section 2.2.8) such as "i-klingon" and "i-bnn". (Other
471 grandfathered tags have a primary language subtag in their first
474 8. Other values MUST NOT be assigned to the primary subtag except by
475 revision or update of this document.
477 Note: For languages that have both an ISO 639-1 two-character code
478 and an ISO 639-2 three-character code, only the ISO 639-1 two-
479 character code is defined in the IANA registry.
481 Note: For languages that have no ISO 639-1 two-character code and for
482 which the ISO 639-2/T (Terminology) code and the ISO 639-2/B
483 (Bibliographic) codes differ, only the Terminology code is defined in
484 the IANA registry. At the time this document was created, all
485 languages that had both kinds of three-character code were also
486 assigned a two-character code; it is not expected that future
487 assignments of this nature will occur.
489 Note: To avoid problems with versioning and subtag choice as
490 experienced during the transition between RFC 1766 and RFC 3066, as
491 well as the canonical nature of subtags defined by this document, the
492 ISO 639 Registration Authority Joint Advisory Committee (ISO 639/
493 RA-JAC) has included the following statement in [iso639.prin]:
495 "A language code already in ISO 639-2 at the point of freezing ISO
496 639-1 shall not later be added to ISO 639-1. This is to ensure
497 consistency in usage over time, since users are directed in Internet
498 applications to employ the alpha-3 code when an alpha-2 code for that
499 language is not available."
506 Phillips & Davis Best Current Practice [Page 9]
508 RFC 4646 Tags for Identifying Languages September 2006
511 In order to avoid instability in the canonical form of tags, if a
512 two-character code is added to ISO 639-1 for a language for which a
513 three-character code was already included in ISO 639-2, the two-
514 character code MUST NOT be registered. See Section 3.4.
516 For example, if some content were tagged with 'haw' (Hawaiian), which
517 currently has no two-character code, the tag would not be invalidated
518 if ISO 639-1 were to assign a two-character code to the Hawaiian
519 language at a later date.
521 For example, one of the grandfathered IANA registrations is
522 "i-enochian". The subtag 'enochian' could be registered in the IANA
523 registry as a primary language subtag (assuming that ISO 639 does not
524 register this language first), making tags such as "enochian-AQ" and
525 "enochian-Latn" valid.
527 2.2.2. Extended Language Subtags
529 The following rules apply to the extended language subtags:
531 1. Three-letter subtags immediately following the primary subtag are
532 reserved for future standardization, anticipating work that is
533 currently under way on ISO 639.
535 2. Extended language subtags MUST follow the primary subtag and
536 precede any other subtags.
538 3. There MAY be up to three extended language subtags.
540 4. Extended language subtags MUST NOT be registered or used to form
541 language tags. Their syntax is described here so that
542 implementations can be compatible with any future revision of
543 this document that does provide for their registration.
545 Extended language subtag records, once they appear in the registry,
546 MUST include exactly one 'Prefix' field indicating an appropriate
547 language subtag or sequence of subtags that MUST always appear as a
548 prefix to the extended language subtag.
550 Example: In a future revision or update of this document, the tag
551 "zh-gan" (registered under RFC 3066) might become a valid non-
552 grandfathered (that is, redundant) tag in which the subtag 'gan'
553 might represent the Chinese dialect 'Gan'.
562 Phillips & Davis Best Current Practice [Page 10]
564 RFC 4646 Tags for Identifying Languages September 2006
569 Script subtags are used to indicate the script or writing system
570 variations that distinguish the written forms of a language or its
571 dialects. The following rules apply to the script subtags:
573 1. All four-character subtags were defined according to
574 [ISO15924]--"Codes for the representation of names of scripts":
575 alpha-4 script codes, or subsequently assigned by the ISO 15924
576 maintenance agency or governing standardization bodies, denoting
577 the script or writing system used in conjunction with this
580 2. Script subtags MUST immediately follow the primary language
581 subtag and all extended language subtags and MUST occur before
582 any other type of subtag described below.
584 3. The script subtags 'Qaaa' through 'Qabx' are reserved for private
585 use in language tags. These subtags correspond to codes reserved
586 by ISO 15924 for private use. These codes MAY be used for non-
587 registered script values. Please refer to Section 4.5 for more
588 information on private use subtags.
590 4. Script subtags MUST NOT be registered using the process in
591 Section 3.5 of this document. Variant subtags MAY be considered
592 for registration for that purpose.
594 5. There MUST be at most one script subtag in a language tag, and
595 the script subtag SHOULD be omitted when it adds no
596 distinguishing value to the tag or when the primary language
597 subtag's record includes a Suppress-Script field listing the
598 applicable script subtag.
600 Example: "sr-Latn" represents Serbian written using the Latin script.
604 Region subtags are used to indicate linguistic variations associated
605 with or appropriate to a specific country, territory, or region.
606 Typically, a region subtag is used to indicate regional dialects or
607 usage, or region-specific spelling conventions. A region subtag can
608 also be used to indicate that content is expressed in a way that is
609 appropriate for use throughout a region, for instance, Spanish
610 content tailored to be useful throughout Latin America.
618 Phillips & Davis Best Current Practice [Page 11]
620 RFC 4646 Tags for Identifying Languages September 2006
623 The following rules apply to the region subtags:
625 1. Region subtags MUST follow any language, extended language, or
626 script subtags and MUST precede all other subtags.
628 2. All two-character subtags following the primary subtag were
629 defined in the IANA registry according to the assignments found
630 in [ISO3166-1] ("Codes for the representation of names of
631 countries and their subdivisions -- Part 1: Country codes") using
632 the list of alpha-2 country codes, or using assignments
633 subsequently made by the ISO 3166 maintenance agency or governing
634 standardization bodies.
636 3. All three-character subtags consisting of digit (numeric)
637 characters following the primary subtag were defined in the IANA
638 registry according to the assignments found in UN Standard
639 Country or Area Codes for Statistical Use [UN_M.49] or
640 assignments subsequently made by the governing standards body.
641 Note that not all of the UN M.49 codes are defined in the IANA
642 registry. The following rules define which codes are entered
643 into the registry as valid subtags:
645 A. UN numeric codes assigned to 'macro-geographical
646 (continental)' or sub-regions MUST be registered in the
647 registry. These codes are not associated with an assigned
648 ISO 3166 alpha-2 code and represent supra-national areas,
649 usually covering more than one nation, state, province, or
652 B. UN numeric codes for 'economic groupings' or 'other
653 groupings' MUST NOT be registered in the IANA registry and
654 MUST NOT be used to form language tags.
656 C. UN numeric codes for countries or areas with ambiguous ISO
657 3166 alpha-2 codes, when entered into the registry, MUST be
658 defined according to the rules in Section 3.4 and MUST be
659 used to form language tags that represent the country or
660 region for which they are defined.
662 D. UN numeric codes for countries or areas for which there is an
663 associated ISO 3166 alpha-2 code in the registry MUST NOT be
664 entered into the registry and MUST NOT be used to form
665 language tags. Note that the ISO 3166-based subtag in the
666 registry MUST actually be associated with the UN M.49 code in
674 Phillips & Davis Best Current Practice [Page 12]
676 RFC 4646 Tags for Identifying Languages September 2006
679 E. UN numeric codes and ISO 3166 alpha-2 codes for countries or
680 areas listed as eligible for registration in [RFC4645] but
681 not presently registered MAY be entered into the IANA
682 registry via the process described in Section 3.5. Once
683 registered, these codes MAY be used to form language tags.
685 F. All other UN numeric codes for countries or areas that do not
686 have an associated ISO 3166 alpha-2 code MUST NOT be entered
687 into the registry and MUST NOT be used to form language tags.
688 For more information about these codes, see Section 3.4.
690 4. Note: The alphanumeric codes in Appendix X of the UN document
691 MUST NOT be entered into the registry and MUST NOT be used to
692 form language tags. (At the time this document was created,
693 these values matched the ISO 3166 alpha-2 codes.)
695 5. There MUST be at most one region subtag in a language tag and the
696 region subtag MAY be omitted, as when it adds no distinguishing
699 6. The region subtags 'AA', 'QM'-'QZ', 'XA'-'XZ', and 'ZZ' are
700 reserved for private use in language tags. These subtags
701 correspond to codes reserved by ISO 3166 for private use. These
702 codes MAY be used for private use region subtags (instead of
703 using a private use subtag sequence). Please refer to
704 Section 4.5 for more information on private use subtags.
706 "de-CH" represents German ('de') as used in Switzerland ('CH').
708 "sr-Latn-CS" represents Serbian ('sr') written using Latin script
709 ('Latn') as used in Serbia and Montenegro ('CS').
711 "es-419" represents Spanish ('es') appropriate to the UN-defined
712 Latin America and Caribbean region ('419').
714 2.2.5. Variant Subtags
716 Variant subtags are used to indicate additional, well-recognized
717 variations that define a language or its dialects that are not
718 covered by other available subtags. The following rules apply to the
721 1. Variant subtags are not associated with any external standard.
722 Variant subtags and their meanings are defined by the
723 registration process defined in Section 3.5.
725 2. Variant subtags MUST follow all of the other defined subtags, but
726 precede any extension or private use subtag sequences.
730 Phillips & Davis Best Current Practice [Page 13]
732 RFC 4646 Tags for Identifying Languages September 2006
735 3. More than one variant MAY be used to form the language tag.
737 4. Variant subtags MUST be registered with IANA according to the
738 rules in Section 3.5 of this document before being used to form
739 language tags. In order to distinguish variants from other types
740 of subtags, registrations MUST meet the following length and
741 content restrictions:
743 1. Variant subtags that begin with a letter (a-z, A-Z) MUST be
744 at least five characters long.
746 2. Variant subtags that begin with a digit (0-9) MUST be at
747 least four characters long.
749 Variant subtag records in the language subtag registry MAY include
750 one or more 'Prefix' fields, which indicate the language tag or tags
751 that would make a suitable prefix (with other subtags, as
752 appropriate) in forming a language tag with the variant. For
753 example, the subtag 'nedis' has a Prefix of "sl", making it suitable
754 to form language tags such as "sl-nedis" and "sl-IT-nedis", but not
755 suitable for use in a tag such as "zh-nedis" or "it-IT-nedis".
757 "sl-nedis" represents the Natisone or Nadiza dialect of Slovenian.
759 "de-CH-1996" represents German as used in Switzerland and as written
760 using the spelling reform beginning in the year 1996 C.E.
762 Most variants that share a prefix are mutually exclusive. For
763 example, the German orthographic variations '1996' and '1901' SHOULD
764 NOT be used in the same tag, as they represent the dates of different
765 spelling reforms. A variant that can meaningfully be used in
766 combination with another variant SHOULD include a 'Prefix' field in
767 its registry record that lists that other variant. For example, if
768 another German variant 'example' were created that made sense to use
769 with '1996', then 'example' should include two Prefix fields: "de"
772 2.2.6. Extension Subtags
774 Extensions provide a mechanism for extending language tags for use in
775 various applications. See Section 3.7. The following rules apply to
778 1. Extension subtags are separated from the other subtags defined
779 in this document by a single-character subtag ("singleton").
780 The singleton MUST be one allocated to a registration authority
781 via the mechanism described in Section 3.7 and MUST NOT be the
782 letter 'x', which is reserved for private use subtag sequences.
786 Phillips & Davis Best Current Practice [Page 14]
788 RFC 4646 Tags for Identifying Languages September 2006
791 2. Note: Private use subtag sequences starting with the singleton
792 subtag 'x' are described in Section 2.2.7 below.
794 3. An extension MUST follow at least a primary language subtag.
795 That is, a language tag cannot begin with an extension.
796 Extensions extend language tags, they do not override or replace
797 them. For example, "a-value" is not a well-formed language tag,
798 while "de-a-value" is.
800 4. Each singleton subtag MUST appear at most one time in each tag
801 (other than as a private use subtag). That is, singleton
802 subtags MUST NOT be repeated. For example, the tag
803 "en-a-bbb-a-ccc" is invalid because the subtag 'a' appears
804 twice. Note that the tag "en-a-bbb-x-a-ccc" is valid because
805 the second appearance of the singleton 'a' is in a private use
808 5. Extension subtags MUST meet all of the requirements for the
809 content and format of subtags defined in this document.
811 6. Extension subtags MUST meet whatever requirements are set by the
812 document that defines their singleton prefix and whatever
813 requirements are provided by the maintaining authority.
815 7. Each extension subtag MUST be from two to eight characters long
816 and consist solely of letters or digits, with each subtag
817 separated by a single '-'.
819 8. Each singleton MUST be followed by at least one extension
820 subtag. For example, the tag "tlh-a-b-foo" is invalid because
821 the first singleton 'a' is followed immediately by another
824 9. Extension subtags MUST follow all language, extended language,
825 script, region, and variant subtags in a tag.
827 10. All subtags following the singleton and before another singleton
828 are part of the extension. Example: In the tag "fr-a-Latn", the
829 subtag 'Latn' does not represent the script subtag 'Latn'
830 defined in the IANA Language Subtag Registry. Its meaning is
831 defined by the extension 'a'.
833 11. In the event that more than one extension appears in a single
834 tag, the tag SHOULD be canonicalized as described in
842 Phillips & Davis Best Current Practice [Page 15]
844 RFC 4646 Tags for Identifying Languages September 2006
847 For example, if the prefix singleton 'r' and the shown subtags were
848 defined, then the following tag would be a valid example:
849 "en-Latn-GB-boont-r-extended-sequence-x-private".
851 2.2.7. Private Use Subtags
853 Private use subtags are used to indicate distinctions in language
854 important in a given context by private agreement. The following
855 rules apply to private use subtags:
857 1. Private use subtags are separated from the other subtags defined
858 in this document by the reserved single-character subtag 'x'.
860 2. Private use subtags MUST conform to the format and content
861 constraints defined in the ABNF for all subtags.
863 3. Private use subtags MUST follow all language, extended language,
864 script, region, variant, and extension subtags in the tag.
865 Another way of saying this is that all subtags following the
866 singleton 'x' MUST be considered private use. Example: The
867 subtag 'US' in the tag "en-x-US" is a private use subtag.
869 4. A tag MAY consist entirely of private use subtags.
871 5. No source is defined for private use subtags. Use of private use
872 subtags is by private agreement only.
874 6. Private use subtags are NOT RECOMMENDED where alternatives exist
875 or for general interchange. See Section 4.5 for more information
876 on private use subtag choice.
878 For example: Users who wished to utilize codes from the Ethnologue
879 publication of SIL International for language identification might
880 agree to exchange tags such as "az-Arab-x-AZE-derbend". This example
881 contains two private use subtags. The first is 'AZE' and the second
884 2.2.8. Preexisting RFC 3066 Registrations
886 Existing IANA-registered language tags from RFC 1766 and/or RFC 3066
887 maintain their validity. These tags will be maintained in the
888 registry in records of either the "grandfathered" or "redundant"
889 type. Grandfathered tags contain one or more subtags that are not
890 defined in the Language Subtag Registry (see Section 3). Redundant
891 tags consist entirely of subtags defined above and whose independent
892 registration is superseded by this document. For more information,
898 Phillips & Davis Best Current Practice [Page 16]
900 RFC 4646 Tags for Identifying Languages September 2006
903 It is important to note that all language tags formed under the
904 guidelines in this document were either legal, well-formed tags or
905 could have been registered under RFC 3066.
907 2.2.9. Classes of Conformance
909 Implementations sometimes need to describe their capabilities with
910 regard to the rules and practices described in this document. There
911 are two classes of conforming implementations described by this
912 document: "well-formed" processors and "validating" processors.
913 Claims of conformance SHOULD explicitly reference one of these
916 An implementation that claims to check for well-formed language tags
919 o Check that the tag and all of its subtags, including extension and
920 private use subtags, conform to the ABNF or that the tag is on the
921 list of grandfathered tags.
923 o Check that singleton subtags that identify extensions do not
924 repeat. For example, the tag "en-a-xx-b-yy-a-zz" is not well-
927 Well-formed processors are strongly encouraged to implement the
928 canonicalization rules contained in Section 4.4.
930 An implementation that claims to be validating MUST:
932 o Check that the tag is well-formed.
934 o Specify the particular registry date for which the implementation
935 performs validation of subtags.
937 o Check that either the tag is a grandfathered tag, or that all
938 language, script, region, and variant subtags consist of valid
939 codes for use in language tags according to the IANA registry as
940 of the particular date specified by the implementation.
942 o Specify which, if any, extension RFCs as defined in Section 3.7
943 are supported, including version, revision, and date.
945 o For any such extensions supported, check that all subtags used in
946 that extension are valid.
948 o For variant and extended language subtags, if the registry
949 contains one or more 'Prefix' fields for that subtag, check that
950 the tag matches at least one prefix. The tag matches if all the
954 Phillips & Davis Best Current Practice [Page 17]
956 RFC 4646 Tags for Identifying Languages September 2006
959 subtags in the 'Prefix' also appear in the tag. For example, the
960 prefix "es-CO" matches the tag "es-Latn-CO-x-private" because both
961 the 'es' language subtag and 'CO' region subtag appear in the tag.
963 3. Registry Format and Maintenance
965 This section defines the Language Subtag Registry and the maintenance
966 and update procedures associated with it, as well as a registry for
967 extensions to language tags (Section 3.7).
969 The Language Subtag Registry contains a comprehensive list of all of
970 the subtags valid in language tags. This allows implementers a
971 straightforward and reliable way to validate language tags. The
972 Language Subtag Registry will be maintained so that, except for
973 extension subtags, it is possible to validate all of the subtags that
974 appear in a language tag under the provisions of this document or its
975 revisions or successors. In addition, the meaning of the various
976 subtags will be unambiguous and stable over time. (The meaning of
977 private use subtags, of course, is not defined by the IANA registry.)
979 3.1. Format of the IANA Language Subtag Registry
981 The IANA Language Subtag Registry ("the registry") consists of a text
982 file that is machine readable in the format described in this
983 section, plus copies of the registration forms approved in accordance
984 with the process described in Section 3.5. The existing registration
985 forms for grandfathered and redundant tags taken from RFC 3066 will
986 be maintained as part of the obsolete RFC 3066 registry. The
987 remaining set of initial subtags will not have registration forms
990 The registry is in the text format described below. This format was
991 based on the record-jar format described in [record-jar].
993 Each line of text is limited to 72 characters, including all
994 whitespace. Records are separated by lines containing only the
995 sequence "%%" (%x25.25).
997 Each field can be viewed as a single, logical line of ASCII
998 characters, comprising a field-name and a field-body separated by a
999 COLON character (%x3A). For convenience, the field-body portion of
1000 this conceptual entity can be split into a multiple-line
1001 representation; this is called "folding". The format of the registry
1002 is described by the following ABNF (per [RFC4234]):
1010 Phillips & Davis Best Current Practice [Page 18]
1012 RFC 4646 Tags for Identifying Languages September 2006
1015 registry = record *("%%" CRLF record)
1016 record = 1*( field-name *SP ":" *SP field-body CRLF )
1017 field-name = (ALPHA / DIGIT) [*(ALPHA / DIGIT / "-") (ALPHA / DIGIT)]
1018 field-body = *(ASCCHAR/LWSP)
1019 ASCCHAR = %x21-25 / %x27-7E / UNICHAR ; Note: AMPERSAND is %x26
1020 UNICHAR = "&#x" 2*6HEXDIG ";"
1022 Figure 2: Registry Format ABNF
1024 The sequence '..' (%x2E.2E) in a field-body denotes a range of
1025 values. Such a range represents all subtags of the same length that
1026 are in alphabetic or numeric order within that range, including the
1027 values explicitly mentioned. For example 'a..c' denotes the values
1028 'a', 'b', and 'c' and '11..13' denotes the values '11', '12', and
1031 Characters from outside the US-ASCII [ISO646] repertoire, as well as
1032 the AMPERSAND character ("&", %x26) when it occurs in a field-body,
1033 are represented by a "Numeric Character Reference" using hexadecimal
1034 notation in the style used by [XML10] (see
1035 <http://www.w3.org/TR/REC-xml/#dt-charref>). This consists of the
1036 sequence "&#x" (%x26.23.78) followed by a hexadecimal representation
1037 of the character's code point in [ISO10646] followed by a closing
1038 semicolon (%x3B). For example, the EURO SIGN, U+20AC, would be
1039 represented by the sequence "€". Note that the hexadecimal
1040 notation MAY have between two and six digits.
1042 All fields whose field-body contains a date value use the "full-date"
1043 format specified in [RFC3339]. For example: "2004-06-28" represents
1044 June 28, 2004, in the Gregorian calendar.
1046 The first record in the file contains the single field whose field-
1047 name is "File-Date" (see Figure 3). The field-body of this record
1048 contains the last modification date of this copy of the registry,
1049 making it possible to compare different versions of the registry.
1050 The registry on the IANA website is the most current. Versions with
1051 an older date than that one are not up-to-date.
1053 File-Date: 2004-06-28
1056 Figure 3: Example of the File-Date Record
1058 Subsequent records represent subtags in the registry. Each of the
1059 fields in each record MUST occur no more than once, unless otherwise
1060 noted below. Each record MUST contain the following fields:
1066 Phillips & Davis Best Current Practice [Page 19]
1068 RFC 4646 Tags for Identifying Languages September 2006
1073 * Type's field-value MUST consist of one of the following
1074 strings: "language", "extlang", "script", "region", "variant",
1075 "grandfathered", and "redundant" and denotes the type of tag or
1078 o Either 'Subtag' or 'Tag'
1080 * Subtag's field-value contains the subtag being defined. This
1081 field MUST only appear in records of whose 'Type' has one of
1082 these values: "language", "extlang", "script", "region", or
1085 * Tag's field-value contains a complete language tag. This field
1086 MUST only appear in records whose 'Type' has one of these
1087 values: "grandfathered" or "redundant". Note that the field-
1088 value will always follow the 'grandfathered' production in the
1093 * Description's field-value contains a non-normative description
1094 of the subtag or tag.
1098 * Added's field-value contains the date the record was added to
1101 The 'Subtag' or 'Tag' field MUST use lowercase letters to form the
1102 subtag or tag, with two exceptions. Subtags whose 'Type' field is
1103 'script' (in other words, subtags defined by ISO 15924) MUST use
1104 titlecase. Subtags whose 'Type' field is 'region' (in other words,
1105 subtags defined by ISO 3166) MUST use uppercase. These exceptions
1106 mirror the use of case in the underlying standards.
1108 The field 'Description' MAY appear more than one time and contains a
1109 description of the tag or subtag in the record. At least one of the
1110 'Description' fields MUST be written or transcribed into the Latin
1111 script; the same or additional fields MAY also include a description
1112 in a non-Latin script. The 'Description' field is used for
1113 identification purposes and SHOULD NOT be taken to represent the
1114 actual native name of the language or variation or to be in any
1115 particular language. Most descriptions are taken directly from
1116 source standards such as ISO 639 or ISO 3166.
1122 Phillips & Davis Best Current Practice [Page 20]
1124 RFC 4646 Tags for Identifying Languages September 2006
1127 Note: Descriptions in registry entries that correspond to ISO 639,
1128 ISO 15924, ISO 3166, or UN M.49 codes are intended only to indicate
1129 the meaning of that identifier as defined in the source standard at
1130 the time it was added to the registry. The description does not
1131 replace the content of the source standard itself. The descriptions
1132 are not intended to be the English localized names for the subtags.
1133 Localization or translation of language tag and subtag descriptions
1134 is out of scope of this document.
1136 Each record MAY also contain the following fields:
1140 * For fields of type 'language', 'extlang', 'script', 'region',
1141 and 'variant', 'Preferred-Value' contains the subtag of the
1142 same 'Type' that is preferred for forming the language tag.
1144 * For fields of type 'grandfathered' and 'redundant', a canonical
1145 mapping to a complete language tag.
1149 * Deprecated's field-value contains the date the record was
1154 * Prefix's field-value contains a language tag with which this
1155 subtag MAY be used to form a new language tag, perhaps with
1156 other subtags as well. This field MUST only appear in records
1157 whose 'Type' field-value is 'variant' or 'extlang'. For
1158 example, the 'Prefix' for the variant 'nedis' is 'sl', meaning
1159 that the tags "sl-nedis" and "sl-IT-nedis" might be appropriate
1160 while the tag "is-nedis" is not.
1164 * Comments contains additional information about the subtag, as
1165 deemed appropriate for understanding the registry and
1166 implementing language tags using the subtag or tag.
1170 * Suppress-Script contains a script subtag that SHOULD NOT be
1171 used to form language tags with the associated primary language
1172 subtag. This field MUST only appear in records whose 'Type'
1173 field-value is 'language'. See Section 4.1.
1178 Phillips & Davis Best Current Practice [Page 21]
1180 RFC 4646 Tags for Identifying Languages September 2006
1183 The field 'Deprecated' MAY be added to any record via the maintenance
1184 process described in Section 3.3 or via the registration process
1185 described in Section 3.5. Usually, the addition of a 'Deprecated'
1186 field is due to the action of one of the standards bodies, such as
1187 ISO 3166, withdrawing a code. In some historical cases, it might not
1188 have been possible to reconstruct the original deprecation date. For
1189 these cases, an approximate date appears in the registry. Although
1190 valid in language tags, subtags and tags with a 'Deprecated' field
1191 are deprecated and validating processors SHOULD NOT generate these
1192 subtags. Note that a record that contains a 'Deprecated' field and
1193 no corresponding 'Preferred-Value' field has no replacement mapping.
1195 The field 'Preferred-Value' contains a mapping between the record in
1196 which it appears and another tag or subtag. The value in this field
1197 is STRONGLY RECOMMENDED as the best choice to represent the value of
1198 this record when selecting a language tag. These values form three
1201 1. ISO 639 language codes that were later withdrawn in favor of
1202 other codes. These values are mostly a historical curiosity.
1204 2. ISO 3166 region codes that have been withdrawn in favor of a new
1205 code. This sometimes happens when a country changes its name or
1206 administration in such a way that warrants a new region code.
1208 3. Tags grandfathered from RFC 3066. In many cases, these tags have
1209 become obsolete because the values they represent were later
1212 Records that contain a 'Preferred-Value' field MUST also have a
1213 'Deprecated' field. This field contains a date of deprecation.
1214 Thus, a language tag processor can use the registry to construct the
1215 valid, non-deprecated set of subtags for a given date. In addition,
1216 for any given tag, a processor can construct the set of valid
1217 language tags that correspond to that tag for all dates up to the
1218 date of the registry. The ability to do these mappings MAY be
1219 beneficial to applications that are matching, selecting, for
1220 filtering content based on its language tags.
1222 Note that 'Preferred-Value' mappings in records of type 'region'
1223 sometimes do not represent exactly the same meaning as the original
1224 value. There are many reasons for a country code to be changed, and
1225 the effect this has on the formation of language tags will depend on
1226 the nature of the change in question.
1228 In particular, the 'Preferred-Value' field does not imply retagging
1229 content that uses the affected subtag.
1234 Phillips & Davis Best Current Practice [Page 22]
1236 RFC 4646 Tags for Identifying Languages September 2006
1239 The field 'Preferred-Value' MUST NOT be modified once created in the
1240 registry. The field MAY be added to records of type "grandfathered"
1241 and "region" according to the rules in Section 3.3. Otherwise the
1242 field MUST NOT be added to any record already in the registry.
1244 The 'Preferred-Value' field in records of type "grandfathered" and
1245 "redundant" contains whole language tags that are strongly
1246 RECOMMENDED for use in place of the record's value. In many cases,
1247 the mappings were created by deprecation of the tags during the
1248 period before this document was adopted. For example, the tag
1249 "no-nyn" was deprecated in favor of the ISO 639-1-defined language
1252 Records of type 'variant' MAY have more than one field of type
1253 'Prefix'. Additional fields of this type MAY be added to a 'variant'
1254 record via the registration process.
1256 Records of type 'extlang' MUST have _exactly_ one 'Prefix' field.
1258 The field-value of the 'Prefix' field consists of a language tag
1259 whose subtags are appropriate to use with this subtag. For example,
1260 the variant subtag '1996' has a 'Prefix' field of "de". This means
1261 that tags starting with the sequence "de-" are appropriate with this
1262 subtag, so "de-Latg-1996" and "de-CH-1996" are both acceptable, while
1263 the tag "fr-1996" is an inappropriate choice.
1265 The field of type 'Prefix' MUST NOT be removed from any record. The
1266 field-value for this type of field MUST NOT be modified.
1268 The field 'Comments' MAY appear more than once per record. This
1269 field MAY be inserted or changed via the registration process and no
1270 guarantee of stability is provided. The content of this field is not
1271 restricted, except by the need to register the information, the
1272 suitability of the request, and by reasonable practical size
1275 The field 'Suppress-Script' MUST only appear in records whose 'Type'
1276 field-value is 'language'. This field MUST NOT appear more than one
1277 time in a record. This field indicates a script used to write the
1278 overwhelming majority of documents for the given language and that
1279 therefore adds no distinguishing information to a language tag. It
1280 helps ensure greater compatibility between the language tags
1281 generated according to the rules in this document and language tags
1282 and tag processors or consumers based on RFC 3066. For example,
1283 virtually all Icelandic documents are written in the Latin script,
1284 making the subtag 'Latn' redundant in the tag "is-Latn".
1290 Phillips & Davis Best Current Practice [Page 23]
1292 RFC 4646 Tags for Identifying Languages September 2006
1295 3.2. Language Subtag Reviewer
1297 The Language Subtag Reviewer is appointed by the IESG for an
1298 indefinite term, subject to removal or replacement at the IESG's
1299 discretion. The Language Subtag Reviewer moderates the ietf-
1300 languages mailing list, responds to requests for registration, and
1301 performs the other registry maintenance duties described in
1302 Section 3.3. Only the Language Subtag Reviewer is permitted to
1303 request IANA to change, update, or add records to the Language Subtag
1306 The performance or decisions of the Language Subtag Reviewer MAY be
1307 appealed to the IESG under the same rules as other IETF decisions
1308 (see [RFC2026]). The IESG can reverse or overturn the decision of
1309 the Language Subtag Reviewer, provide guidance, or take other
1310 appropriate actions.
1312 3.3. Maintenance of the Registry
1314 Maintenance of the registry requires that as codes are assigned or
1315 withdrawn by ISO 639, ISO 15924, ISO 3166, and UN M.49, the Language
1316 Subtag Reviewer MUST evaluate each change, determine whether it
1317 conflicts with existing registry entries, and submit the information
1318 to IANA for inclusion in the registry. If a change takes place and
1319 the Language Subtag Reviewer does not do this in a timely manner,
1320 then any interested party MAY use the procedure in Section 3.5 to
1321 register the appropriate update.
1323 Note: The redundant and grandfathered entries together are the
1324 complete list of tags registered under [RFC3066]. The redundant tags
1325 are those that can now be formed using the subtags defined in the
1326 registry together with the rules of Section 2.2. The grandfathered
1327 entries include those that can never be legal under those same
1330 The set of redundant and grandfathered tags is permanent and stable:
1331 new entries in this section MUST NOT be added and existing entries
1332 MUST NOT be removed. Records of type 'grandfathered' MAY have their
1333 type converted to 'redundant'; see item 12 in Section 3.6 for more
1334 information. The decision-making process about which tags were
1335 initially grandfathered and which were made redundant is described in
1338 RFC 3066 tags that were deprecated prior to the adoption of this
1339 document are part of the list of grandfathered tags, and their
1340 component subtags were not included as registered variants (although
1341 they remain eligible for registration). For example, the tag
1342 "art-lojban" was deprecated in favor of the language subtag 'jbo'.
1346 Phillips & Davis Best Current Practice [Page 24]
1348 RFC 4646 Tags for Identifying Languages September 2006
1351 The Language Subtag Reviewer MUST ensure that new subtags meet the
1352 requirements in Section 4.1 or submit an appropriate alternate subtag
1353 as described in that section. When either a change or addition to
1354 the registry is needed, the Language Subtag Reviewer MUST prepare the
1355 complete record, including all fields, and forward it to IANA for
1356 insertion into the registry. Each record being modified or inserted
1357 MUST be forwarded in a separate message.
1359 If a record represents a new subtag that does not currently exist in
1360 the registry, then the message's subject line MUST include the word
1361 "INSERT". If the record represents a change to an existing subtag,
1362 then the subject line of the message MUST include the word "MODIFY".
1363 The message MUST contain both the record for the subtag being
1364 inserted or modified and the new File-Date record. Here is an
1365 example of what the body of the message might contain:
1367 LANGUAGE SUBTAG MODIFICATION
1368 File-Date: 2005-01-02
1372 Description: Natisone dialect
1373 Description: Nadiza dialect
1376 Comments: This is a comment shown
1380 Figure 4: Example of a Language Subtag Modification Form
1382 Whenever an entry is created or modified in the registry, the
1383 'File-Date' record at the start of the registry is updated to reflect
1384 the most recent modification date in the [RFC3339] "full-date"
1387 Before forwarding a new registration to IANA, the Language Subtag
1388 Reviewer MUST ensure that values in the 'Subtag' field match case
1389 according to the description in Section 3.1.
1391 3.4. Stability of IANA Registry Entries
1393 The stability of entries and their meaning in the registry is
1394 critical to the long-term stability of language tags. The rules in
1395 this section guarantee that a specific language tag's meaning is
1396 stable over time and will not change.
1402 Phillips & Davis Best Current Practice [Page 25]
1404 RFC 4646 Tags for Identifying Languages September 2006
1407 These rules specifically deal with how changes to codes (including
1408 withdrawal and deprecation of codes) maintained by ISO 639, ISO
1409 15924, ISO 3166, and UN M.49 are reflected in the IANA Language
1410 Subtag Registry. Assignments to the IANA Language Subtag Registry
1411 MUST follow the following stability rules:
1413 1. Values in the fields 'Type', 'Subtag', 'Tag', 'Added',
1414 'Deprecated' and 'Preferred-Value' MUST NOT be changed and are
1415 guaranteed to be stable over time.
1417 2. Values in the 'Description' field MUST NOT be changed in a way
1418 that would invalidate previously-existing tags. They MAY be
1419 broadened somewhat in scope, changed to add information, or
1420 adapted to the most common modern usage. For example, countries
1421 occasionally change their official names; a historical example
1422 of this would be "Upper Volta" changing to "Burkina Faso".
1424 3. Values in the field 'Prefix' MAY be added to records of type
1425 'variant' via the registration process.
1427 4. Values in the field 'Prefix' MAY be modified, so long as the
1428 modifications broaden the set of prefixes. That is, a prefix
1429 MAY be replaced by one of its own prefixes. For example, the
1430 prefix "en-US" could be replaced by "en", but not by the
1431 prefixes "en-Latn", "fr", or "en-US-boont". If one of those
1432 prefixes were needed, a new Prefix SHOULD be registered.
1434 5. Values in the field 'Prefix' MUST NOT be removed.
1436 6. The field 'Comments' MAY be added, changed, modified, or removed
1437 via the registration process or any of the processes or
1438 considerations described in this section.
1440 7. The field 'Suppress-Script' MAY be added or removed via the
1441 registration process.
1443 8. Codes assigned by ISO 639, ISO 15924, and ISO 3166 that do not
1444 conflict with existing subtags of the associated type and whose
1445 meaning is not the same as an existing subtag of the same type
1446 are entered into the IANA registry as new records.
1448 9. Codes assigned by ISO 639, ISO 15924, or ISO 3166 that are
1449 withdrawn by their respective maintenance or registration
1450 authority remain valid in language tags. A 'Deprecated' field
1451 containing the date of withdrawal is added to the record. If a
1452 new record of the same type is added that represents a
1458 Phillips & Davis Best Current Practice [Page 26]
1460 RFC 4646 Tags for Identifying Languages September 2006
1463 replacement value, then a 'Preferred-Value' field MAY also be
1464 added. The registration process MAY be used to add comments
1465 about the withdrawal of the code by the respective standard.
1468 The region code 'TL' was assigned to the country 'Timor-
1469 Leste', replacing the code 'TP' (which was assigned to 'East
1470 Timor' when it was under administration by Portugal). The
1471 subtag 'TP' remains valid in language tags, but its record
1472 contains the a 'Preferred-Value' of 'TL' and its field
1473 'Deprecated' contains the date the new code was assigned
1476 10. Codes assigned by ISO 639, ISO 15924, or ISO 3166 that conflict
1477 with existing subtags of the associated type, including subtags
1478 that are deprecated, MUST NOT be entered into the registry. The
1479 following additional considerations apply to subtag values that
1482 A. For ISO 639 codes, if the newly assigned code's meaning is
1483 not represented by a subtag in the IANA registry, the
1484 Language Subtag Reviewer, as described in Section 3.5, SHALL
1485 prepare a proposal for entering in the IANA registry as soon
1486 as practical a registered language subtag as an alternate
1487 value for the new code. The form of the registered language
1488 subtag will be at the discretion of the Language Subtag
1489 Reviewer and MUST conform to other restrictions on language
1490 subtags in this document.
1492 B. For all subtags whose meaning is derived from an external
1493 standard (i.e., ISO 639, ISO 15924, ISO 3166, or UN M.49),
1494 if a new meaning is assigned to an existing code and the new
1495 meaning broadens the meaning of that code, then the meaning
1496 for the associated subtag MAY be changed to match. The
1497 meaning of a subtag MUST NOT be narrowed, however, as this
1498 can result in an unknown proportion of the existing uses of
1499 a subtag becoming invalid. Note: ISO 639 maintenance
1500 agency/registration authority (MA/RA) has adopted a similar
1503 C. For ISO 15924 codes, if the newly assigned code's meaning is
1504 not represented by a subtag in the IANA registry, the
1505 Language Subtag Reviewer, as described in Section 3.5, SHALL
1506 prepare a proposal for entering in the IANA registry as soon
1507 as practical a registered variant subtag as an alternate
1508 value for the new code. The form of the registered variant
1514 Phillips & Davis Best Current Practice [Page 27]
1516 RFC 4646 Tags for Identifying Languages September 2006
1519 subtag will be at the discretion of the Language Subtag
1520 Reviewer and MUST conform to other restrictions on variant
1521 subtags in this document.
1523 D. For ISO 3166 codes, if the newly assigned code's meaning is
1524 associated with the same UN M.49 code as another 'region'
1525 subtag, then the existing region subtag remains as the
1526 preferred value for that region and no new entry is created.
1527 A comment MAY be added to the existing region subtag
1528 indicating the relationship to the new ISO 3166 code.
1530 E. For ISO 3166 codes, if the newly assigned code's meaning is
1531 associated with a UN M.49 code that is not represented by an
1532 existing region subtag, then the Language Subtag Reviewer,
1533 as described in Section 3.5, SHALL prepare a proposal for
1534 entering the appropriate UN M.49 country code as an entry in
1537 F. For ISO 3166 codes, if there is no associated UN numeric
1538 code, then the Language Subtag Reviewer SHALL petition the
1539 UN to create one. If there is no response from the UN
1540 within ninety days of the request being sent, the Language
1541 Subtag Reviewer SHALL prepare a proposal for entering in the
1542 IANA registry as soon as practical a registered variant
1543 subtag as an alternate value for the new code. The form of
1544 the registered variant subtag will be at the discretion of
1545 the Language Subtag Reviewer and MUST conform to other
1546 restrictions on variant subtags in this document. This
1547 situation is very unlikely to ever occur.
1549 11. UN M.49 has codes for both countries and areas (such as '276'
1550 for Germany) and geographical regions and sub-regions (such as
1551 '150' for Europe). UN M.49 country or area codes for which
1552 there is no corresponding ISO 3166 code SHOULD NOT be
1553 registered, except as a surrogate for an ISO 3166 code that is
1554 blocked from registration by an existing subtag. If such a code
1555 becomes necessary, then the registration authority for ISO 3166
1556 SHOULD first be petitioned to assign a code to the region. If
1557 the petition for a code assignment by ISO 3166 is refused or not
1558 acted on in a timely manner, the registration process described
1559 in Section 3.5 MAY then be used to register the corresponding UN
1560 M.49 code. At the time this document was written, there were
1561 only four such codes: 830 (Channel Islands), 831 (Guernsey), 832
1562 (Jersey), and 833 (Isle of Man). This way, UN M.49 codes remain
1563 available as the value of last resort in cases where ISO 3166
1564 reassigns a deprecated value in the registry.
1570 Phillips & Davis Best Current Practice [Page 28]
1572 RFC 4646 Tags for Identifying Languages September 2006
1575 12. Stability provisions apply to grandfathered tags with this
1576 exception: should all of the subtags in a grandfathered tag
1577 become valid subtags in the IANA registry, then the field 'Type'
1578 in that record is changed from 'grandfathered' to 'redundant'.
1579 Note that this will not affect language tags that match the
1580 grandfathered tag, since these tags will now match valid
1581 generative subtag sequences. For example, if the subtag 'gan'
1582 in the language tag "zh-gan" were to be registered as an
1583 extended language subtag, then the grandfathered tag "zh-gan"
1584 would be deprecated (but existing content or implementations
1585 that use "zh-gan" would remain valid).
1587 3.5. Registration Procedure for Subtags
1589 The procedure given here MUST be used by anyone who wants to use a
1590 subtag not currently in the IANA Language Subtag Registry.
1592 Only subtags of type 'language' and 'variant' will be considered for
1593 independent registration of new subtags. Handling of subtags needed
1594 for stability and subtags necessary to keep the registry synchronized
1595 with ISO 639, ISO 15924, ISO 3166, and UN M.49 within the limits
1596 defined by this document are described in Section 3.3. Stability
1597 provisions are described in Section 3.4.
1599 This procedure MAY also be used to register or alter the information
1600 for the 'Description', 'Comments', 'Deprecated', or 'Prefix' fields
1601 in a subtag's record as described in Section 3.4. Changes to all
1602 other fields in the IANA registry are NOT permitted.
1604 Registering a new subtag or requesting modifications to an existing
1605 tag or subtag starts with the requester filling out the registration
1606 form reproduced below. Note that each response is not limited in
1607 size so that the request can adequately describe the registration.
1608 The fields in the "Record Requested" section SHOULD follow the
1609 requirements in Section 3.1.
1626 Phillips & Davis Best Current Practice [Page 29]
1628 RFC 4646 Tags for Identifying Languages September 2006
1631 LANGUAGE SUBTAG REGISTRATION FORM
1632 1. Name of requester:
1633 2. E-mail address of requester:
1634 3. Record Requested:
1645 4. Intended meaning of the subtag:
1646 5. Reference to published description
1647 of the language (book or article):
1648 6. Any other relevant information:
1650 Figure 5: The Language Subtag Registration Form
1652 The subtag registration form MUST be sent to
1653 <ietf-languages@iana.org> for a two-week review period before it can
1654 be submitted to IANA. (This is an open list and can be joined by
1655 sending a request to <ietf-languages-request@iana.org>.)
1657 Variant subtags are usually registered for use with a particular
1658 range of language tags. For example, the subtag 'rozaj' is intended
1659 for use with language tags that start with the primary language
1660 subtag "sl", since Resian is a dialect of Slovenian. Thus, the
1661 subtag 'rozaj' would be appropriate in tags such as "sl-Latn-rozaj"
1662 or "sl-IT-rozaj". This information is stored in the 'Prefix' field
1663 in the registry. Variant registration requests SHOULD include at
1664 least one 'Prefix' field in the registration form.
1666 Extended language subtags are reserved for future standardization.
1667 These subtags will be REQUIRED to include exactly one 'Prefix' field
1668 once they are allowed for registration.
1670 The 'Prefix' field for a given registered subtag exists in the IANA
1671 registry as a guide to usage. Additional prefixes MAY be added by
1672 filing an additional registration form. In that form, the "Any other
1673 relevant information:" field MUST indicate that it is the addition of
1676 Requests to add a prefix to a variant subtag that imply a different
1677 semantic meaning will probably be rejected. For example, a request
1678 to add the prefix "de" to the subtag 'nedis' so that the tag
1682 Phillips & Davis Best Current Practice [Page 30]
1684 RFC 4646 Tags for Identifying Languages September 2006
1687 "de-nedis" represented some German dialect would be rejected. The
1688 'nedis' subtag represents a particular Slovenian dialect and the
1689 additional registration would change the semantic meaning assigned to
1690 the subtag. A separate subtag SHOULD be proposed instead.
1692 The 'Description' field MUST contain a description of the tag being
1693 registered written or transcribed into the Latin script; it MAY also
1694 include a description in a non-Latin script. Non-ASCII characters
1695 MUST be escaped using the syntax described in Section 3.1. The
1696 'Description' field is used for identification purposes and doesn't
1697 necessarily represent the actual native name of the language or
1698 variation or to be in any particular language.
1700 While the 'Description' field itself is not guaranteed to be stable
1701 and errata corrections MAY be undertaken from time to time, attempts
1702 to provide translations or transcriptions of entries in the registry
1703 itself will probably be frowned upon by the community or rejected
1704 outright, as changes of this nature have an impact on the provisions
1707 When the two-week period has passed, the Language Subtag Reviewer
1708 either forwards the record to be inserted or modified to
1709 iana@iana.org according to the procedure described in Section 3.3, or
1710 rejects the request because of significant objections raised on the
1711 list or due to problems with constraints in this document (which MUST
1712 be explicitly cited). The Language Subtag Reviewer MAY also extend
1713 the review period in two-week increments to permit further
1714 discussion. The Language Subtag Reviewer MUST indicate on the list
1715 whether the registration has been accepted, rejected, or extended
1716 following each two-week period.
1718 Note that the Language Subtag Reviewer MAY raise objections on the
1719 list if he or she so desires. The important thing is that the
1720 objection MUST be made publicly.
1722 The applicant is free to modify a rejected application with
1723 additional information and submit it again; this restarts the two-
1724 week comment period.
1726 Decisions made by the Language Subtag Reviewer MAY be appealed to the
1727 IESG [RFC2028] under the same rules as other IETF decisions
1730 All approved registration forms are available online in the directory
1731 http://www.iana.org/numbers.html under "languages".
1738 Phillips & Davis Best Current Practice [Page 31]
1740 RFC 4646 Tags for Identifying Languages September 2006
1743 Updates or changes to existing records follow the same procedure as
1744 new registrations. The Language Subtag Reviewer decides whether
1745 there is consensus to update the registration following the two-week
1746 review period; normally, objections by the original registrant will
1747 carry extra weight in forming such a consensus.
1749 Registrations are permanent and stable. Once registered, subtags
1750 will not be removed from the registry and will remain a valid way in
1751 which to specify a specific language or variant.
1753 Note: The purpose of the "Description" in the registration form is to
1754 aid people trying to verify whether a language is registered or what
1755 language or language variation a particular subtag refers to. In
1756 most cases, reference to an authoritative grammar or dictionary of
1757 that language will be useful; in cases where no such work exists,
1758 other well-known works describing that language or in that language
1759 MAY be appropriate. The Language Subtag Reviewer decides what
1760 constitutes "good enough" reference material. This requirement is
1761 not intended to exclude particular languages or dialects due to the
1762 size of the speaker population or lack of a standardized orthography.
1763 Minority languages will be considered equally on their own merits.
1765 3.6. Possibilities for Registration
1767 Possibilities for registration of subtags or information about
1770 o Primary language subtags for languages not listed in ISO 639 that
1771 are not variants of any listed or registered language MAY be
1772 registered. At the time this document was created, there were no
1773 examples of this form of subtag. Before attempting to register a
1774 language subtag, there MUST be an attempt to register the language
1775 with ISO 639. Subtags MUST NOT be registered for codes that exist
1776 in ISO 639-1 or ISO 639-2, that are under consideration by the ISO
1777 639 maintenance or registration authorities, or that have never
1778 been attempted for registration with those authorities. If ISO
1779 639 has previously rejected a language for registration, it is
1780 reasonable to assume that there must be additional, very
1781 compelling evidence of need before it will be registered in the
1782 IANA registry (to the extent that it is very unlikely that any
1783 subtags will be registered of this type).
1785 o Dialect or other divisions or variations within a language, its
1786 orthography, writing system, regional or historical usage,
1787 transliteration or other transformation, or distinguishing
1788 variation MAY be registered as variant subtags. An example is the
1789 'rozaj' subtag (the Resian dialect of Slovenian).
1794 Phillips & Davis Best Current Practice [Page 32]
1796 RFC 4646 Tags for Identifying Languages September 2006
1799 o The addition or maintenance of fields (generally of an
1800 informational nature) in Tag or Subtag records as described in
1801 Section 3.1 and subject to the stability provisions in
1802 Section 3.4. This includes descriptions, comments, deprecation
1803 and preferred values for obsolete or withdrawn codes, or the
1804 addition of script or extlang information to primary language
1807 o The addition of records and related field value changes necessary
1808 to reflect assignments made by ISO 639, ISO 15924, ISO 3166, and
1809 UN M.49 as described in Section 3.4.
1811 Subtags proposed for registration that would cause all or part of a
1812 grandfathered tag to become redundant but whose meaning conflicts
1813 with or alters the meaning of the grandfathered tag MUST be rejected.
1815 This document leaves the decision on what subtags or changes to
1816 subtags are appropriate (or not) to the registration process
1817 described in Section 3.5.
1819 Note: four-character primary language subtags are reserved to allow
1820 for the possibility of alpha4 codes in some future addition to the
1821 ISO 639 family of standards.
1823 ISO 639 defines a maintenance agency for additions to and changes in
1824 the list of languages in ISO 639. This agency is:
1826 International Information Centre for Terminology (Infoterm)
1827 Aichholzgasse 6/12, AT-1120
1829 Phone: +43 1 26 75 35 Ext. 312 Fax: +43 1 216 32 72
1831 ISO 639-2 defines a maintenance agency for additions to and changes
1832 in the list of languages in ISO 639-2. This agency is:
1835 Network Development and MARC Standards Office
1836 Washington, D.C. 20540 USA
1837 Phone: +1 202 707 6237 Fax: +1 202 707 0115
1838 URL: http://www.loc.gov/standards/iso639-2
1850 Phillips & Davis Best Current Practice [Page 33]
1852 RFC 4646 Tags for Identifying Languages September 2006
1855 The maintenance agency for ISO 3166 (country codes) is:
1857 ISO 3166 Maintenance Agency
1858 c/o International Organization for Standardization
1860 CH-1211 Geneva 20 Switzerland
1861 Phone: +41 22 749 72 33 Fax: +41 22 749 73 49
1862 URL: http://www.iso.org/iso/en/prods-services/iso3166ma/index.html
1864 The registration authority for ISO 15924 (script codes) is:
1866 Unicode Consortium Box 391476
1867 Mountain View, CA 94039-1476, USA
1868 URL: http://www.unicode.org/iso15924
1870 The Statistics Division of the United Nations Secretariat maintains
1871 the Standard Country or Area Codes for Statistical Use and can be
1874 Statistical Services Branch
1876 United Nations, Room DC2-1620
1877 New York, NY 10017, USA
1879 Fax: +1-212-963-0623
1880 E-mail: statistics@un.org
1881 URL: http://unstats.un.org/unsd/methods/m49/m49alpha.htm
1883 3.7. Extensions and Extensions Registry
1885 Extension subtags are those introduced by single-character subtags
1886 ("singletons") other than 'x'. They are reserved for the generation
1887 of identifiers that contain a language component and are compatible
1888 with applications that understand language tags.
1890 The structure and form of extensions are defined by this document so
1891 that implementations can be created that are forward compatible with
1892 applications that might be created using singletons in the future.
1893 In addition, defining a mechanism for maintaining singletons will
1894 lend stability to this document by reducing the likely need for
1895 future revisions or updates.
1897 Single-character subtags are assigned by IANA using the "IETF
1898 Consensus" policy defined by [RFC2434]. This policy requires the
1899 development of an RFC, which SHALL define the name, purpose,
1900 processes, and procedures for maintaining the subtags. The
1901 maintaining or registering authority, including name, contact email,
1906 Phillips & Davis Best Current Practice [Page 34]
1908 RFC 4646 Tags for Identifying Languages September 2006
1911 discussion list email, and URL location of the registry, MUST be
1912 indicated clearly in the RFC. The RFC MUST specify or include each
1915 o The specification MUST reference the specific version or revision
1916 of this document that governs its creation and MUST reference this
1917 section of this document.
1919 o The specification and all subtags defined by the specification
1920 MUST follow the ABNF and other rules for the formation of tags and
1921 subtags as defined in this document. In particular, it MUST
1922 specify that case is not significant and that subtags MUST NOT
1923 exceed eight characters in length.
1925 o The specification MUST specify a canonical representation.
1927 o The specification of valid subtags MUST be available over the
1928 Internet and at no cost.
1930 o The specification MUST be in the public domain or available via a
1931 royalty-free license acceptable to the IETF and specified in the
1934 o The specification MUST be versioned, and each version of the
1935 specification MUST be numbered, dated, and stable.
1937 o The specification MUST be stable. That is, extension subtags,
1938 once defined by a specification, MUST NOT be retracted or change
1939 in meaning in any substantial way.
1941 o The specification MUST include in a separate section the
1942 registration form reproduced in this section (below) to be used in
1943 registering the extension upon publication as an RFC.
1945 o IANA MUST be informed of changes to the contact information and
1946 URL for the specification.
1948 IANA will maintain a registry of allocated single-character
1949 (singleton) subtags. This registry MUST use the record-jar format
1950 described by the ABNF in Section 3.1. Upon publication of an
1951 extension as an RFC, the maintaining authority defined in the RFC
1952 MUST forward this registration form to iesg@ietf.org, who MUST
1953 forward the request to iana@iana.org. The maintaining authority of
1954 the extension MUST maintain the accuracy of the record by sending an
1955 updated full copy of the record to iana@iana.org with the subject
1956 line "LANGUAGE TAG EXTENSION UPDATE" whenever content changes. Only
1957 the 'Comments', 'Contact_Email', 'Mailing_List', and 'URL' fields MAY
1958 be modified in these updates.
1962 Phillips & Davis Best Current Practice [Page 35]
1964 RFC 4646 Tags for Identifying Languages September 2006
1967 Failure to maintain this record, maintain the corresponding registry,
1968 or meet other conditions imposed by this section of this document MAY
1969 be appealed to the IESG [RFC2028] under the same rules as other IETF
1970 decisions (see [RFC2026]) and MAY result in the authority to maintain
1971 the extension being withdrawn or reassigned by the IESG.
1985 Figure 6: Format of Records in the Language Tag Extensions Registry
1987 'Identifier' contains the single-character subtag (singleton)
1988 assigned to the extension. The Internet-Draft submitted to define
1989 the extension SHOULD specify which letter or digit to use, although
1990 the IESG MAY change the assignment when approving the RFC.
1992 'Description' contains the name and description of the extension.
1994 'Comments' is an OPTIONAL field and MAY contain a broader description
1997 'Added' contains the date the RFC was published in the "full-date"
1998 format specified in [RFC3339]. For example: 2004-06-28 represents
1999 June 28, 2004, in the Gregorian calendar.
2001 'RFC' contains the RFC number assigned to the extension.
2003 'Authority' contains the name of the maintaining authority for the
2006 'Contact_Email' contains the email address used to contact the
2007 maintaining authority.
2009 'Mailing_List' contains the URL or subscription email address of the
2010 mailing list used by the maintaining authority.
2012 'URL' contains the URL of the registry for this extension.
2018 Phillips & Davis Best Current Practice [Page 36]
2020 RFC 4646 Tags for Identifying Languages September 2006
2023 The determination of whether an Internet-Draft meets the above
2024 conditions and the decision to grant or withhold such authority rests
2025 solely with the IESG and is subject to the normal review and appeals
2026 process associated with the RFC process.
2028 Extension authors are strongly cautioned that many (including most
2029 well-formed) processors will be unaware of any special relationships
2030 or meaning inherent in the order of extension subtags. Extension
2031 authors SHOULD avoid subtag relationships or canonicalization
2032 mechanisms that interfere with matching or with length restrictions
2033 that sometimes exist in common protocols where the extension is used.
2034 In particular, applications MAY truncate the subtags in doing
2035 matching or in fitting into limited lengths, so it is RECOMMENDED
2036 that the most significant information be in the most significant
2037 (left-most) subtags and that the specification gracefully handle
2040 When a language tag is to be used in a specific, known, protocol, it
2041 is RECOMMENDED that the language tag not contain extensions not
2042 supported by that protocol. In addition, note that some protocols
2043 MAY impose upper limits on the length of the strings used to store or
2044 transport the language tag.
2046 3.8. Initialization of the Registries
2048 Upon adoption of this document, an initial version of the Language
2049 Subtag Registry containing the various subtags initially valid in a
2050 language tag is necessary. This collection of subtags, along with a
2051 description of the process used to create it, is described by
2052 [RFC4645]. IANA SHALL publish the initial version of the registry
2053 described by this document from the content of [RFC4645]. Once
2054 published by IANA, the maintenance procedures, rules, and
2055 registration processes described in this document will be available
2056 for new registrations or updates.
2058 Registrations that are in process under the rules defined in
2059 [RFC3066] when this document is adopted MAY be completed under the
2060 former rules, at the discretion of the Language Tag Reviewer (as
2061 described in [RFC3066]). Until the IESG officially appoints a
2062 Language Subtag Reviewer, the existing Language Tag Reviewer SHALL
2063 serve as the Language Subtag Reviewer.
2065 Any new registrations submitted using the RFC 3066 forms or format
2066 after the adoption of this document and publication of the registry
2067 by IANA MUST be rejected.
2074 Phillips & Davis Best Current Practice [Page 37]
2076 RFC 4646 Tags for Identifying Languages September 2006
2079 An initial version of the Language Tag Extensions Registry described
2080 in Section 3.7 is also needed. The Language Tag Extensions Registry
2081 SHALL be initialized with a single record containing a single field
2082 of type "File-Date" as a placeholder for future assignments.
2084 4. Formation and Processing of Language Tags
2086 This section addresses how to use the information in the registry
2087 with the tag syntax to choose, form, and process language tags.
2089 4.1. Choice of Language Tag
2091 One is sometimes faced with the choice between several possible tags
2092 for the same body of text.
2094 Interoperability is best served when all users use the same language
2095 tag in order to represent the same language. If an application has
2096 requirements that make the rules here inapplicable, then that
2097 application risks damaging interoperability. It is strongly
2098 RECOMMENDED that users not define their own rules for language tag
2101 Subtags SHOULD only be used where they add useful distinguishing
2102 information; extraneous subtags interfere with the meaning,
2103 understanding, and processing of language tags. In particular, users
2104 and implementations SHOULD follow the 'Prefix' and 'Suppress-Script'
2105 fields in the registry (defined in Section 3.1): these fields provide
2106 guidance on when specific additional subtags SHOULD (and SHOULD NOT)
2107 be used in a language tag.
2109 Of particular note, many applications can benefit from the use of
2110 script subtags in language tags, as long as the use is consistent for
2111 a given context. Script subtags were not formally defined in RFC
2112 3066 and their use can affect matching and subtag identification by
2113 implementations of RFC 3066, as these subtags appear between the
2114 primary language and region subtags. For example, if a user requests
2115 content in an implementation of Section 2.5 of [RFC3066] using the
2116 language range "en-US", content labeled "en-Latn-US" will not match
2117 the request. Therefore, it is important to know when script subtags
2118 will customarily be used and when they ought not be used. In the
2119 registry, the Suppress-Script field helps ensure greater
2120 compatibility between the language tags generated according to the
2121 rules in this document and language tags and tag processors or
2122 consumers based on RFC 3066 by defining when users SHOULD NOT include
2123 a script subtag with a particular primary language subtag.
2130 Phillips & Davis Best Current Practice [Page 38]
2132 RFC 4646 Tags for Identifying Languages September 2006
2135 Extended language subtags (type 'extlang' in the registry; see
2136 Section 3.1) also appear between the primary language and region
2137 subtags and are reserved for future standardization. Applications
2138 might benefit from their judicious use in forming language tags in
2139 the future. Similar recommendations are expected to apply to their
2140 use as apply to script subtags.
2142 Standards, protocols, and applications that reference this document
2143 normatively but apply different rules to the ones given in this
2144 section MUST specify how the procedure varies from the one given
2147 The choice of subtags used to form a language tag SHOULD be guided by
2148 the following rules:
2150 1. Use as precise a tag as possible, but no more specific than is
2151 justified. Avoid using subtags that are not important for
2152 distinguishing content in an application.
2154 * For example, 'de' might suffice for tagging an email written
2155 in German, while "de-CH-1996" is probably unnecessarily
2156 precise for such a task.
2158 2. The script subtag SHOULD NOT be used to form language tags unless
2159 the script adds some distinguishing information to the tag. The
2160 field 'Suppress-Script' in the primary language record in the
2161 registry indicates which script subtags do not add distinguishing
2162 information for most applications.
2164 * For example, the subtag 'Latn' should not be used with the
2165 primary language 'en' because nearly all English documents are
2166 written in the Latin script and it adds no distinguishing
2167 information. However, if a document were written in English
2168 mixing Latin script with another script such as Braille
2169 ('Brai'), then it might be appropriate to choose to indicate
2170 both scripts to aid in content selection, such as the
2171 application of a style sheet.
2173 3. If a tag or subtag has a 'Preferred-Value' field in its registry
2174 entry, then the value of that field SHOULD be used to form the
2175 language tag in preference to the tag or subtag in which the
2176 preferred value appears.
2178 * For example, use 'he' for Hebrew in preference to 'iw'.
2186 Phillips & Davis Best Current Practice [Page 39]
2188 RFC 4646 Tags for Identifying Languages September 2006
2191 4. The 'und' (Undetermined) primary language subtag SHOULD NOT be
2192 used to label content, even if the language is unknown. Omitting
2193 the language tag altogether is preferred to using a tag with a
2194 primary language subtag of 'und'. The 'und' subtag MAY be useful
2195 for protocols that require a language tag to be provided. The
2196 'und' subtag MAY also be useful when matching language tags in
2199 5. The 'mul' (Multiple) primary language subtag SHOULD NOT be used
2200 whenever the protocol allows the separate tags for multiple
2201 languages, as is the case for the Content-Language header in
2202 HTTP. The 'mul' subtag conveys little useful information:
2203 content in multiple languages SHOULD individually tag the
2204 languages where they appear or otherwise indicate the actual
2205 language in preference to the 'mul' subtag.
2207 6. The same variant subtag SHOULD NOT be used more than once within
2210 * For example, do not use "de-DE-1901-1901".
2212 To ensure consistent backward compatibility, this document contains
2213 several provisions to account for potential instability in the
2214 standards used to define the subtags that make up language tags.
2215 These provisions mean that no language tag created under the rules in
2216 this document will become obsolete.
2218 4.2. Meaning of the Language Tag
2220 The relationship between the tag and the information it relates to is
2221 defined by the context in which the tag appears. Accordingly, this
2222 section gives only possible examples of its usage.
2224 o For a single information object, the associated language tags
2225 might be interpreted as the set of languages that is necessary for
2226 a complete comprehension of the complete object. Example: Plain
2229 o For an aggregation of information objects, the associated language
2230 tags could be taken as the set of languages used inside components
2231 of that aggregation. Examples: Document stores and libraries.
2233 o For information objects whose purpose is to provide alternatives,
2234 the associated language tags could be regarded as a hint that the
2235 content is provided in several languages and that one has to
2236 inspect each of the alternatives in order to find its language or
2237 languages. In this case, the presence of multiple tags might not
2238 mean that one needs to be multi-lingual to get complete
2242 Phillips & Davis Best Current Practice [Page 40]
2244 RFC 4646 Tags for Identifying Languages September 2006
2247 understanding of the document. Example: MIME multipart/
2250 o In markup languages, such as HTML and XML, language information
2251 can be added to each part of the document identified by the markup
2252 structure (including the whole document itself). For example, one
2253 could write <span lang="fr">C'est la vie.</span> inside a
2254 Norwegian document; the Norwegian-speaking user could then access
2255 a French-Norwegian dictionary to find out what the marked section
2256 meant. If the user were listening to that document through a
2257 speech synthesis interface, this formation could be used to signal
2258 the synthesizer to appropriately apply French text-to-speech
2259 pronunciation rules to that span of text, instead of applying the
2260 inappropriate Norwegian rules.
2262 Language tags are related when they contain a similar sequence of
2263 subtags. For example, if a language tag B contains language tag A as
2264 a prefix, then B is typically "narrower" or "more specific" than A.
2265 Thus, "zh-Hant-TW" is more specific than "zh-Hant".
2267 This relationship is not guaranteed in all cases: specifically,
2268 languages that begin with the same sequence of subtags are NOT
2269 guaranteed to be mutually intelligible, although they might be. For
2270 example, the tag "az" shares a prefix with both "az-Latn"
2271 (Azerbaijani written using the Latin script) and "az-Cyrl"
2272 (Azerbaijani written using the Cyrillic script). A person fluent in
2273 one script might not be able to read the other, even though the text
2274 might be identical. Content tagged as "az" most probably is written
2275 in just one script and thus might not be intelligible to a reader
2276 familiar with the other script.
2278 4.3. Length Considerations
2280 [RFC3066] did not provide an upper limit on the size of language
2281 tags. While RFC 3066 did define the semantics of particular subtags
2282 in such a way that most language tags consisted of language and
2283 region subtags with a combined total length of up to six characters,
2284 larger registered tags were not only possible but were actually
2287 Neither the language tag syntax nor other requirements in this
2288 document impose a fixed upper limit on the number of subtags in a
2289 language tag (and thus an upper bound on the size of a tag). The
2290 language tag syntax suggests that, depending on the specific
2291 language, more subtags (and thus a longer tag) are sometimes
2292 necessary to completely identify the language for certain
2293 applications; thus, it is possible to envision long or complex subtag
2298 Phillips & Davis Best Current Practice [Page 41]
2300 RFC 4646 Tags for Identifying Languages September 2006
2303 4.3.1. Working with Limited Buffer Sizes
2305 Some applications and protocols are forced to allocate fixed buffer
2306 sizes or otherwise limit the length of a language tag. A conformant
2307 implementation or specification MAY refuse to support the storage of
2308 language tags that exceed a specified length. Any such limitation
2309 SHOULD be clearly documented, and such documentation SHOULD include
2310 what happens to longer tags (for example, whether an error value is
2311 generated or the language tag is truncated). A protocol that allows
2312 tags to be truncated at an arbitrary limit, without giving any
2313 indication of what that limit is, has the potential for causing harm
2314 by changing the meaning of tags in substantial ways.
2316 In practice, most language tags do not require more than a few
2317 subtags and will not approach reasonably sized buffer limitations;
2320 Some specifications or protocols have limits on tag length but do not
2321 have a fixed length limitation. For example, [RFC2231] has no
2322 explicit length limitation: the length available for the language tag
2323 is constrained by the length of other header components (such as the
2324 charset's name) coupled with the 76-character limit in [RFC2047].
2325 Thus, the "limit" might be 50 or more characters, but it could
2326 potentially be quite small.
2328 The considerations for assigning a buffer limit are:
2330 Implementations SHOULD NOT truncate language tags unless the
2331 meaning of the tag is purposefully being changed, or unless the
2332 tag does not fit into a limited buffer size specified by a
2333 protocol for storage or transmission.
2335 Implementations SHOULD warn the user when a tag is truncated since
2336 truncation changes the semantic meaning of the tag.
2338 Implementations of protocols or specifications that are space
2339 constrained but do not have a fixed limit SHOULD use the longest
2340 possible tag in preference to truncation.
2342 Protocols or specifications that specify limited buffer sizes for
2343 language tags MUST allow for language tags of up to 33 characters.
2345 Protocols or specifications that specify limited buffer sizes for
2346 language tags SHOULD allow for language tags of at least 42
2354 Phillips & Davis Best Current Practice [Page 42]
2356 RFC 4646 Tags for Identifying Languages September 2006
2359 The following illustration shows how the 42-character recommendation
2360 was derived. The combination of language and extended language
2361 subtags was chosen for future compatibility. At up to 15 characters,
2362 this combination is longer than the longest possible primary language
2363 subtag (8 characters):
2365 language = 3 (ISO 639-2; ISO 639-1 requires 2)
2366 extlang1 = 4 (each subsequent subtag includes '-')
2367 extlang2 = 4 (unlikely: needs prefix="language-extlang1")
2368 extlang3 = 4 (extremely unlikely)
2369 script = 5 (if not suppressed: see Section 4.1)
2370 region = 4 (UN M.49; ISO 3166 requires 3)
2371 variant1 = 9 (MUST have language as a prefix)
2372 variant2 = 9 (MUST have language-variant1 as a prefix)
2374 total = 42 characters
2376 Figure 7: Derivation of the Limit on Tag Length
2378 4.3.2. Truncation of Language Tags
2380 Truncation of a language tag alters the meaning of the tag, and thus
2381 SHOULD be avoided. However, truncation of language tags is sometimes
2382 necessary due to limited buffer sizes. Such truncation MUST NOT
2383 permit a subtag to be chopped off in the middle or the formation of
2384 invalid tags (for example, one ending with the "-" character).
2386 This means that applications or protocols that truncate tags MUST do
2387 so by progressively removing subtags along with their preceding "-"
2388 from the right side of the language tag until the tag is short enough
2389 for the given buffer. If the resulting tag ends with a single-
2390 character subtag, that subtag and its preceding "-" MUST also be
2391 removed. For example:
2393 Tag to truncate: zh-Latn-CN-variant1-a-extend1-x-wadegile-private1
2394 1. zh-Latn-CN-variant1-a-extend1-x-wadegile
2395 2. zh-Latn-CN-variant1-a-extend1
2396 3. zh-Latn-CN-variant1
2401 Figure 8: Example of Tag Truncation
2410 Phillips & Davis Best Current Practice [Page 43]
2412 RFC 4646 Tags for Identifying Languages September 2006
2415 4.4. Canonicalization of Language Tags
2417 Since a particular language tag is sometimes used by many processes,
2418 language tags SHOULD always be created or generated in a canonical
2421 A language tag is in canonical form when:
2423 1. The tag is well-formed according the rules in Section 2.1 and
2426 2. Subtags of type 'Region' that have a Preferred-Value mapping in
2427 the IANA registry (see Section 3.1) SHOULD be replaced with their
2428 mapped value. Note: In rare cases, the mapped value will also
2429 have a Preferred-Value.
2431 3. Redundant or grandfathered tags that have a Preferred-Value
2432 mapping in the IANA registry (see Section 3.1) MUST be replaced
2433 with their mapped value. These items either are deprecated
2434 mappings created before the adoption of this document (such as
2435 the mapping of "no-nyn" to "nn" or "i-klingon" to "tlh") or are
2436 the result of later registrations or additions to this document
2437 (for example, "zh-guoyu" might be mapped to a language-extlang
2438 combination such as "zh-cmn" by some future update of this
2441 4. Other subtags that have a Preferred-Value mapping in the IANA
2442 registry (see Section 3.1) MUST be replaced with their mapped
2443 value. These items consist entirely of clerical corrections to
2444 ISO 639-1 in which the deprecated subtags have been maintained
2445 for compatibility purposes.
2447 5. If more than one extension subtag sequence exists, the extension
2448 sequences are ordered into case-insensitive ASCII order by
2451 Example: The language tag "en-A-aaa-B-ccc-bbb-x-xyz" is in canonical
2452 form, while "en-B-ccc-bbb-A-aaa-X-xyz" is well-formed but not in
2455 Example: The language tag "en-BU" (English as used in Burma) is not
2456 canonical because the 'BU' subtag has a canonical mapping to 'MM'
2457 (Myanmar), although the tag "en-BU" maintains its validity.
2459 Canonicalization of language tags does not imply anything about the
2460 use of upper or lowercase letters when processing or comparing
2461 subtags (and as described in Section 2.1). All comparisons MUST be
2462 performed in a case-insensitive manner.
2466 Phillips & Davis Best Current Practice [Page 44]
2468 RFC 4646 Tags for Identifying Languages September 2006
2471 When performing canonicalization of language tags, processors MAY
2472 regularize the case of the subtags (that is, this process is
2473 OPTIONAL), following the case used in the registry. Note that this
2474 corresponds to the following casing rules: uppercase all non-initial
2475 two-letter subtags; titlecase all non-initial four-letter subtags;
2476 lowercase everything else.
2478 Note: Case folding of ASCII letters in certain locales, unless
2479 carefully handled, sometimes produces non-ASCII character values.
2480 The Unicode Character Database file "SpecialCasing.txt" defines the
2481 specific cases that are known to cause problems with this. In
2482 particular, the letter 'i' (U+0069) in Turkish and Azerbaijani is
2483 uppercased to U+0130 (LATIN CAPITAL LETTER I WITH DOT ABOVE).
2484 Implementers SHOULD specify a locale-neutral casing operation to
2485 ensure that case folding of subtags does not produce this value,
2486 which is illegal in language tags. For example, if one were to
2487 uppercase the region subtag 'in' using Turkish locale rules, the
2488 sequence U+0130 U+004E would result instead of the expected 'IN'.
2490 Note: if the field 'Deprecated' appears in a registry record without
2491 an accompanying 'Preferred-Value' field, then that tag or subtag is
2492 deprecated without a replacement. Validating processors SHOULD NOT
2493 generate tags that include these values, although the values are
2494 canonical when they appear in a language tag.
2496 An extension MUST define any relationships that exist between the
2497 various subtags in the extension and thus MAY define an alternate
2498 canonicalization scheme for the extension's subtags. Extensions MAY
2499 define how the order of the extension's subtags are interpreted. For
2500 example, an extension could define that its subtags are in canonical
2501 order when the subtags are placed into ASCII order: that is,
2502 "en-a-aaa-bbb-ccc" instead of "en-a-ccc-bbb-aaa". Another extension
2503 might define that the order of the subtags influences their semantic
2504 meaning (so that "en-b-ccc-bbb-aaa" has a different value from
2505 "en-b-aaa-bbb-ccc"). However, extension specifications SHOULD be
2506 designed so that they are tolerant of the typical processes described
2509 4.5. Considerations for Private Use Subtags
2511 Private use subtags, like all other subtags, MUST conform to the
2512 format and content constraints in the ABNF. Private use subtags have
2513 no meaning outside the private agreement between the parties that
2514 intend to use or exchange language tags that employ them. The same
2515 subtags MAY be used with a different meaning under a separate private
2516 agreement. They SHOULD NOT be used where alternatives exist and
2517 SHOULD NOT be used in content or protocols intended for general use.
2522 Phillips & Davis Best Current Practice [Page 45]
2524 RFC 4646 Tags for Identifying Languages September 2006
2527 Private use subtags are simply useless for information exchange
2528 without prior arrangement. The value and semantic meaning of private
2529 use tags and of the subtags used within such a language tag are not
2530 defined by this document.
2532 Subtags defined in the IANA registry as having a specific private use
2533 meaning convey more information that a purely private use tag
2534 prefixed by the singleton subtag 'x'. For applications, this
2535 additional information MAY be useful.
2537 For example, the region subtags 'AA', 'ZZ', and in the ranges
2538 'QM'-'QZ' and 'XA'-'XZ' (derived from ISO 3166 private use codes) MAY
2539 be used to form a language tag. A tag such as "zh-Hans-XQ" conveys a
2540 great deal of public, interchangeable information about the language
2541 material (that it is Chinese in the simplified Chinese script and is
2542 suitable for some geographic region 'XQ'). While the precise
2543 geographic region is not known outside of private agreement, the tag
2544 conveys far more information than an opaque tag such as "x-someLang",
2545 which contains no information about the language subtag or script
2546 subtag outside of the private agreement.
2548 However, in some cases content tagged with private use subtags MAY
2549 interact with other systems in a different and possibly unsuitable
2550 manner compared to tags that use opaque, privately defined subtags,
2551 so the choice of the best approach sometimes depends on the
2552 particular domain in question.
2554 5. IANA Considerations
2556 This section deals with the processes and requirements necessary for
2557 IANA to undertake to maintain the subtag and extension registries as
2558 defined by this document and in accordance with the requirements of
2561 The impact on the IANA maintainers of the two registries defined by
2562 this document will be a small increase in the frequency of new
2565 5.1. Language Subtag Registry
2567 Upon adoption of this document, the registry will be initialized by a
2568 companion document: [RFC4645]. The criteria and process for
2569 selecting the initial set of records are described in that document.
2570 The initial set of records represents no impact on IANA, since the
2571 work to create it will be performed externally.
2578 Phillips & Davis Best Current Practice [Page 46]
2580 RFC 4646 Tags for Identifying Languages September 2006
2583 The new registry MUST be listed under "Language Tags" at
2584 <http://www.iana.org/numbers.html>, replacing the existing
2585 registrations defined by [RFC3066]. The existing set of registration
2586 forms and RFC 3066 registrations MUST be relabeled as "Language Tags
2587 (Obsolete)" and maintained (but not added to or modified).
2589 Future work on the Language Subtag Registry SHALL be limited to
2590 inserting or replacing whole records preformatted for IANA by the
2591 Language Subtag Reviewer as described in Section 3.3 of this document
2592 and archiving the forwarded registration form.
2594 Each record MUST be sent to iana@iana.org with a subject line
2595 indicating whether the enclosed record is an insertion of a new
2596 record (indicated by the word "INSERT" in the subject line) or a
2597 replacement of an existing record (indicated by the word "MODIFY" in
2598 the subject line). Records MUST NOT be deleted from the registry.
2599 IANA MUST place any inserted or modified records into the appropriate
2600 section of the language subtag registry, grouping the records by
2601 their 'Type' field. Inserted records MAY be placed anywhere in the
2602 appropriate section; there is no guarantee of the order of the
2603 records beyond grouping them together by 'Type'. Modified records
2604 MUST overwrite the record they replace.
2606 Included in any request to insert or modify records MUST be a new
2607 File-Date record. This record MUST be placed first in the registry.
2608 In the event that the File-Date record present in the registry has a
2609 later date than the record being inserted or modified, the existing
2610 record MUST be preserved.
2612 5.2. Extensions Registry
2614 The Language Tag Extensions Registry will also be generated and sent
2615 to IANA as described in Section 3.7. This registry can contain at
2616 most 35 records, and thus changes to this registry are expected to be
2619 Future work by IANA on the Language Tag Extensions Registry is
2620 limited to two cases. First, the IESG MAY request that new records
2621 be inserted into this registry from time to time. These requests
2622 MUST include the record to insert in the exact format described in
2623 Section 3.7. In addition, there MAY be occasional requests from the
2624 maintaining authority for a specific extension to update the contact
2625 information or URLs in the record. These requests MUST include the
2626 complete, updated record. IANA is not responsible for validating the
2627 information provided, only that it is properly formatted. It should
2628 reasonably be seen to come from the maintaining authority named in
2629 the record present in the registry.
2634 Phillips & Davis Best Current Practice [Page 47]
2636 RFC 4646 Tags for Identifying Languages September 2006
2639 6. Security Considerations
2641 Language tags used in content negotiation, like any other information
2642 exchanged on the Internet, might be a source of concern because they
2643 might be used to infer the nationality of the sender, and thus
2644 identify potential targets for surveillance.
2646 This is a special case of the general problem that anything sent is
2647 visible to the receiving party and possibly to third parties as well.
2648 It is useful to be aware that such concerns can exist in some cases.
2650 The evaluation of the exact magnitude of the threat, and any possible
2651 countermeasures, is left to each application protocol (see BCP 72
2652 [RFC3552] for best current practice guidance on security threats and
2655 The language tag associated with a particular information item is of
2656 no consequence whatsoever in determining whether that content might
2657 contain possible homographs. The fact that a text is tagged as being
2658 in one language or using a particular script subtag provides no
2659 assurance whatsoever that it does not contain characters from scripts
2660 other than the one(s) associated with or specified by that language
2663 Since there is no limit to the number of variant, private use, and
2664 extension subtags, and consequently no limit on the possible length
2665 of a tag, implementations need to guard against buffer overflow
2666 attacks. See Section 4.3 for details on language tag truncation,
2667 which can occur as a consequence of defenses against buffer overflow.
2669 Although the specification of valid subtags for an extension (see
2670 Section 3.7) MUST be available over the Internet, implementations
2671 SHOULD NOT mechanically depend on it being always accessible, to
2672 prevent denial-of-service attacks.
2674 7. Character Set Considerations
2676 The syntax in this document requires that language tags use only the
2677 characters A-Z, a-z, 0-9, and HYPHEN-MINUS, which are present in most
2678 character sets, so the composition of language tags should not have
2679 any character set issues.
2681 Rendering of characters based on the content of a language tag is not
2682 addressed in this memo. Historically, some languages have relied on
2683 the use of specific character sets or other information in order to
2684 infer how a specific character should be rendered (notably this
2685 applies to language- and culture-specific variations of Han
2686 ideographs as used in Japanese, Chinese, and Korean). When language
2690 Phillips & Davis Best Current Practice [Page 48]
2692 RFC 4646 Tags for Identifying Languages September 2006
2695 tags are applied to spans of text, rendering engines sometimes use
2696 that information in deciding which font to use in the absence of
2697 other information, particularly where languages with distinct writing
2698 traditions use the same characters.
2700 8. Changes from RFC 3066
2702 The main goals for this revision of language tags were the following:
2704 *Compatibility.* All RFC 3066 language tags (including those in the
2705 IANA registry) remain valid in this specification. The changes in
2706 this document represent additional constraints on language tags.
2707 That is, in no case is the syntax more permissive and processors
2708 based on the ABNF and other provisions of RFC 3066 (such as those
2709 described in [XMLSchema]) will be able to process the tags described
2710 by this document. In addition, this document defines language tags
2711 in such as way as to ensure future compatibility.
2713 *Stability.* Because of changes in the past in the underlying ISO
2714 standards, a valid RFC 3066 language tag could become invalid or have
2715 its meaning change. This has the potential of invalidating content
2716 that may have an extensive shelf-life. In this specification, once a
2717 language tag is valid, it remains valid forever.
2719 *Validity.* The structure of language tags defined by this document
2720 makes it possible to determine if a particular tag is well-formed
2721 without regard for the actual content or "meaning" of the tag as a
2722 whole. This is important because the registry grows and underlying
2723 standards change over time. In addition, it must be possible to
2724 determine if a tag is valid (or not) for a given point in time in
2725 order to provide reproducible, testable results. This process must
2726 not be error-prone; otherwise implementations might give different
2727 results. By having an authoritative registry with specific
2728 versioning information, the validity of language tags at any point in
2729 time can be precisely determined (instead of interpolating values
2730 from many separate sources).
2732 *Utility.* It is sometimes important to be able to differentiate
2733 between written forms of a language -- for many implementations this
2734 is more important than distinguishing between the spoken variants of
2735 a language. Languages are written in a wide variety of different
2736 scripts, so this document provides for the generative use of ISO
2737 15924 script codes. Like the generative use of ISO language and
2738 country codes in RFC 3066, this allows combinations to be produced
2739 without resorting to the registration process. The addition of UN
2740 M.49 codes provides for the generation of language tags with regional
2741 scope, which is also required by some applications.
2746 Phillips & Davis Best Current Practice [Page 49]
2748 RFC 4646 Tags for Identifying Languages September 2006
2751 The recast of the registry from containing whole language tags to
2752 subtags is a key part of this. An important feature of RFC 3066 was
2753 that it allowed generative use of subtags. This allows people to
2754 meaningfully use generated tags, without the delays in registering
2755 whole tags or the need to register all of the combinations that might
2758 The choice of placing the extended language and script subtags
2759 between the primary language and region subtags was widely debated.
2760 This design was chosen because the prevalent matching and content
2761 negotiation schemes rely on the subtags being arranged in order of
2762 increasing specificity. That is, the subtags that mark a greater
2763 barrier to mutual intelligibility appear left-most in a tag. For
2764 example, when selecting content written in Azerbaijani, the script
2765 (Arabic, Cyrillic, or Latin) represents a greater barrier to
2766 understanding than any regional variations (those associated with
2767 Azerbaijan or Iran, for example). Individuals who prefer documents
2768 in a particular script, but can deal with the minor regional
2769 differences, can therefore select appropriate content. Applications
2770 that do not deal with written content will continue to omit these
2773 *Extensibility.* Because of the widespread use of language tags, it
2774 is disruptive to have periodic revisions of the core specification,
2775 even in the face of demonstrated need. The extension mechanism
2776 provides for a way for independent RFCs to define extensions to
2777 language tags. These extensions have a very constrained, well-
2778 defined structure that prevents extensions from interfering with
2779 implementations of language tags defined in this document.
2781 The document also anticipates features of ISO 639-3 with the addition
2782 of the extended language subtags, as well as the possibility of other
2783 ISO 639 parts becoming useful for the formation of language tags in
2786 The use and definition of private use tags have also been modified,
2787 to allow people to use private use subtags to extend or modify
2788 defined tags and to move as much information as possible out of
2789 private use and into the regular structure.
2791 The goal for each of these modifications is to reduce or eliminate
2792 the need for future revisions of this document.
2802 Phillips & Davis Best Current Practice [Page 50]
2804 RFC 4646 Tags for Identifying Languages September 2006
2807 The specific changes in this document to meet these goals are:
2809 o Defines the ABNF and rules for subtags so that the category of all
2810 subtags can be determined without reference to the registry.
2812 o Adds the concept of well-formed vs. validating processors,
2813 defining the rules by which an implementation can claim to be one
2816 o Replaces the IANA language tag registry with a language subtag
2817 registry that provides a complete list of valid subtags in the
2818 IANA registry. This allows for robust implementation and ease of
2819 maintenance. The language subtag registry becomes the canonical
2820 source for forming language tags.
2822 o Provides a process that guarantees stability of language tags, by
2823 handling reuse of values by ISO 639, ISO 15924, and ISO 3166 in
2824 the event that they register a previously used value for a new
2827 o Allows ISO 15924 script code subtags and allows them to be used
2828 generatively. Defines a method for indicating in the registry
2829 when script subtags are necessary for a given language tag.
2831 o Adds the concept of a variant subtag and allows variants to be
2834 o Adds the ability to use a class of UN M.49 tags for supra-national
2835 regions and to resolve conflicts in the assignment of ISO 3166
2838 o Defines the private use tags in ISO 639, ISO 15924, and ISO 3166
2839 as the mechanism for creating private use language, script, and
2840 region subtags, respectively.
2842 o Adds a well-defined extension mechanism.
2844 o Defines an extended language subtag, possibly for use with certain
2845 anticipated features of ISO 639-3.
2858 Phillips & Davis Best Current Practice [Page 51]
2860 RFC 4646 Tags for Identifying Languages September 2006
2865 9.1. Normative References
2867 [ISO10646] International Organization for Standardization,
2868 "ISO/IEC 10646:2003. Information technology --
2869 Universal Multiple-Octet Coded Character Set (UCS)",
2872 [ISO15924] International Organization for Standardization, "ISO
2873 15924:2004. Information and documentation -- Codes for
2874 the representation of names of scripts", January 2004.
2876 [ISO3166-1] International Organization for Standardization, "ISO
2877 3166-1:1997. Codes for the representation of names of
2878 countries and their subdivisions -- Part 1: Country
2881 [ISO639-1] International Organization for Standardization, "ISO
2882 639-1:2002. Codes for the representation of names of
2883 languages -- Part 1: Alpha-2 code", 2002.
2885 [ISO639-2] International Organization for Standardization, "ISO
2886 639-2:1998. Codes for the representation of names of
2887 languages -- Part 2: Alpha-3 code, first edition",
2890 [ISO646] International Organization for Standardization,
2891 "ISO/IEC 646:1991, Information technology -- ISO 7-bit
2892 coded character set for information interchange.",
2895 [RFC2026] Bradner, S., "The Internet Standards Process --
2896 Revision 3", BCP 9, RFC 2026, October 1996.
2898 [RFC2028] Hovey, R. and S. Bradner, "The Organizations Involved
2899 in the IETF Standards Process", BCP 11, RFC 2028,
2902 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate
2903 Requirement Levels", BCP 14, RFC 2119, March 1997.
2905 [RFC2434] Narten, T. and H. Alvestrand, "Guidelines for Writing
2906 an IANA Considerations Section in RFCs", BCP 26,
2907 RFC 2434, October 1998.
2914 Phillips & Davis Best Current Practice [Page 52]
2916 RFC 4646 Tags for Identifying Languages September 2006
2919 [RFC2860] Carpenter, B., Baker, F., and M. Roberts, "Memorandum
2920 of Understanding Concerning the Technical Work of the
2921 Internet Assigned Numbers Authority", RFC 2860,
2924 [RFC3339] Klyne, G., Ed. and C. Newman, "Date and Time on the
2925 Internet: Timestamps", RFC 3339, July 2002.
2927 [RFC4234] Crocker, D., Ed. and P. Overell, "Augmented BNF for
2928 Syntax Specifications: ABNF", RFC 4234, October 2005.
2930 [UN_M.49] Statistics Division, United Nations, "Standard Country
2931 or Area Codes for Statistical Use", UN Standard
2932 Country or Area Codes for Statistical Use, Revision 4
2933 (United Nations publication, Sales No. 98.XVII.9,
2936 9.2. Informative References
2938 [RFC1766] Alvestrand, H., "Tags for the Identification of
2939 Languages", RFC 1766, March 1995.
2941 [RFC2047] Moore, K., "MIME (Multipurpose Internet Mail
2942 Extensions) Part Three: Message Header Extensions for
2943 Non-ASCII Text", RFC 2047, November 1996.
2945 [RFC2231] Freed, N. and K. Moore, "MIME Parameter Value and
2946 Encoded Word Extensions: Character Sets, Languages,
2947 and Continuations", RFC 2231, November 1997.
2949 [RFC2781] Hoffman, P. and F. Yergeau, "UTF-16, an encoding of
2950 ISO 10646", RFC 2781, February 2000.
2952 [RFC3066] Alvestrand, H., "Tags for the Identification of
2953 Languages", BCP 47, RFC 3066, January 2001.
2955 [RFC3552] Rescorla, E. and B. Korver, "Guidelines for Writing
2956 RFC Text on Security Considerations", BCP 72,
2957 RFC 3552, July 2003.
2959 [RFC4645] Ewell, D., Ed., "Initial Language Subtag Registry",
2960 RFC 4645, September 2006.
2962 [RFC4647] Phillips, A., Ed. and M. Davis, Ed., "Matching of
2963 Language Tags", BCP 47, RFC 4647, September 2006.
2970 Phillips & Davis Best Current Practice [Page 53]
2972 RFC 4646 Tags for Identifying Languages September 2006
2975 [Unicode] Unicode Consortium, "The Unicode Standard, Version
2976 5.0", Boston, MA, Addison-Wesley, 2007. ISBN 0-321-
2979 [XML10] Bray (et al), T., "Extensible Markup Language (XML)
2982 [XMLSchema] Biron, P., Ed. and A. Malhotra, Ed., "XML Schema Part
2983 2: Datatypes Second Edition", 10 2004, <
2984 http://www.w3.org/TR/xmlschema-2/>.
2986 [iso639.prin] ISO 639 Joint Advisory Committee, "ISO 639 Joint
2987 Advisory Committee: Working principles for ISO 639
2988 maintenance", March 2000, <http://www.loc.gov/
2989 standards/iso639-2/iso639jac_n3r.html>.
2991 [record-jar] Raymond, E., "The Art of Unix Programming", 2003,
2992 <urn:isbn:0-13-142901-9>.
3026 Phillips & Davis Best Current Practice [Page 54]
3028 RFC 4646 Tags for Identifying Languages September 2006
3031 Appendix A. Acknowledgements
3033 Any list of contributors is bound to be incomplete; please regard the
3034 following as only a selection from the group of people who have
3035 contributed to make this document what it is today.
3037 The contributors to RFC 3066 and RFC 1766, the precursors of this
3038 document, made enormous contributions directly or indirectly to this
3039 document and are generally responsible for the success of language
3042 The following people (in alphabetical order) contributed to this
3043 document or to RFCs 1766 and 3066:
3045 Glenn Adams, Harald Tveit Alvestrand, Tim Berners-Lee, Marc Blanchet,
3046 Nathaniel Borenstein, Karen Broome, Eric Brunner, Sean M. Burke, M.T.
3047 Carrasco Benitez, Jeremy Carroll, John Clews, Jim Conklin, Peter
3048 Constable, John Cowan, Mark Crispin, Dave Crocker, Elwyn Davies,
3049 Martin Duerst, Frank Ellerman, Michael Everson, Doug Ewell, Ned
3050 Freed, Tim Goodwin, Dirk-Willem van Gulik, Marion Gunn, Joel Halpren,
3051 Elliotte Rusty Harold, Paul Hoffman, Scott Hollenbeck, Richard
3052 Ishida, Olle Jarnefors, Kent Karlsson, John Klensin, Erkki
3053 Kolehmainen, Alain LaBonte, Eric Mader, Ira McDonald, Keith Moore,
3054 Chris Newman, Masataka Ohta, Dylan Pierce, Randy Presuhn, George
3055 Rhoten, Felix Sasaki, Markus Scherer, Keld Jorn Simonsen, Thierry
3056 Sourbier, Otto Stolz, Tex Texin, Andrea Vine, Rhys Weatherley, Misha
3057 Wolf, Francois Yergeau and many, many others.
3059 Very special thanks must go to Harald Tveit Alvestrand, who
3060 originated RFCs 1766 and 3066, and without whom this document would
3061 not have been possible. Special thanks must go to Michael Everson,
3062 who has served as Language Tag Reviewer for almost the complete
3063 period since the publication of RFC 1766. Special thanks to Doug
3064 Ewell, for his production of the first complete subtag registry, and
3065 his work in producing a test parser for verifying language tags.
3082 Phillips & Davis Best Current Practice [Page 55]
3084 RFC 4646 Tags for Identifying Languages September 2006
3087 Appendix B. Examples of Language Tags (Informative)
3089 Simple language subtag:
3097 i-enochian (example of a grandfathered tag)
3099 Language subtag plus Script subtag:
3101 zh-Hant (Chinese written using the Traditional Chinese script)
3103 zh-Hans (Chinese written using the Simplified Chinese script)
3105 sr-Cyrl (Serbian written using the Cyrillic script)
3107 sr-Latn (Serbian written using the Latin script)
3109 Language-Script-Region:
3111 zh-Hans-CN (Chinese written using the Simplified script as used in
3114 sr-Latn-CS (Serbian written using the Latin script as used in
3115 Serbia and Montenegro)
3119 sl-rozaj (Resian dialect of Slovenian
3121 sl-nedis (Nadiza dialect of Slovenian)
3123 Language-Region-Variant:
3125 de-CH-1901 (German as used in Switzerland using the 1901 variant
3128 sl-IT-nedis (Slovenian as used in Italy, Nadiza dialect)
3138 Phillips & Davis Best Current Practice [Page 56]
3140 RFC 4646 Tags for Identifying Languages September 2006
3143 Language-Script-Region-Variant:
3145 sl-Latn-IT-nedis (Nadiza dialect of Slovenian written using the
3146 Latin script as used in Italy. Note that this tag is NOT
3147 RECOMMENDED because subtag 'sl' has a Suppress-Script value of
3152 de-DE (German for Germany)
3154 en-US (English as used in the United States)
3156 es-419 (Spanish appropriate for the Latin America and Caribbean
3157 region using the UN region code)
3159 Private use subtags:
3163 az-Arab-x-AZE-derbend
3165 Extended language subtags (examples ONLY: extended languages MUST be
3166 defined by revision or update to this document):
3172 Private use registry values:
3174 x-whatever (private use using the singleton 'x')
3176 qaa-Qaaa-QM-x-southern (all private tags)
3178 de-Qaaa (German, with a private script)
3180 sr-Latn-QM (Serbian, Latin-script, private region)
3182 sr-Qaaa-CS (Serbian, private script, for Serbia and Montenegro)
3184 Tags that use extensions (examples ONLY: extensions MUST be defined
3185 by revision or update to this document or by RFC):
3189 zh-CN-a-myExt-x-private
3194 Phillips & Davis Best Current Practice [Page 57]
3196 RFC 4646 Tags for Identifying Languages September 2006
3199 en-a-myExt-b-another
3203 de-419-DE (two region tags)
3205 a-DE (use of a single-character subtag in primary position; note
3206 that there are a few grandfathered tags that start with "i-" that
3209 ar-a-aaa-b-bbb-a-ccc (two extensions with same single-letter
3214 Addison Phillips (Editor)
3217 EMail: addison@inter-locale.com
3223 EMail: mark.davis@macchiato.com or mark.davis@google.com
3250 Phillips & Davis Best Current Practice [Page 58]
3252 RFC 4646 Tags for Identifying Languages September 2006
3255 Full Copyright Statement
3257 Copyright (C) The Internet Society (2006).
3259 This document is subject to the rights, licenses and restrictions
3260 contained in BCP 78, and except as set forth therein, the authors
3261 retain all their rights.
3263 This document and the information contained herein are provided on an
3264 "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS
3265 OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET
3266 ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED,
3267 INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE
3268 INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED
3269 WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
3271 Intellectual Property
3273 The IETF takes no position regarding the validity or scope of any
3274 Intellectual Property Rights or other rights that might be claimed to
3275 pertain to the implementation or use of the technology described in
3276 this document or the extent to which any license under such rights
3277 might or might not be available; nor does it represent that it has
3278 made any independent effort to identify any such rights. Information
3279 on the procedures with respect to rights in RFC documents can be
3280 found in BCP 78 and BCP 79.
3282 Copies of IPR disclosures made to the IETF Secretariat and any
3283 assurances of licenses to be made available, or the result of an
3284 attempt made to obtain a general license or permission for the use of
3285 such proprietary rights by implementers or users of this
3286 specification can be obtained from the IETF on-line IPR repository at
3287 http://www.ietf.org/ipr.
3289 The IETF invites any interested party to bring to its attention any
3290 copyrights, patents or patent applications, or other proprietary
3291 rights that may cover technology that may be required to implement
3292 this standard. Please address the information to the IETF at
3297 Funding for the RFC Editor function is provided by the IETF
3298 Administrative Support Activity (IASA).
3306 Phillips & Davis Best Current Practice [Page 59]