RFC 3866

Language Tags and Ranges in the Lightweight Directory Access Protocol (LDAP)

📅 published July 2004, from ietf

Language Tags and Ranges in the Lightweight Directory Access Protocol (LDAP) 💖

Okay bestie, this one is by K. Zeilenga, Ed. at the OpenLDAP Foundation, dated July 2004 🎀. It's a Standards Track doc, and it obsoletes RFC 2596, so it's the new version that replaces the old one ✨. The Network Working Group published it as Request for Comments 3866 🦋.

Status of this Memo ✨

This document specifies an Internet standards track protocol for the Internet community, and requests discussion and suggestions for improvements. Please refer to the current edition of the "Internet Official Protocol Standards" (STD 1) for the standardization state and status of this protocol. Distribution of this memo is unlimited.

Copyright Notice 🌸

Abstract 🫶

It is often desirable to be able to indicate the natural language associated with values held in a directory and to be able to query the directory for values which fulfill the user's language needs. This document details the use of Language Tags and Ranges in the Lightweight Directory Access Protocol (LDAP).

1. Background and Intended Use 🧚

LDAP [RFC3377] is how clients poke at and change information stored in a distributed directory system 💅. The info lives as attributes of entries. Most of those attributes are human-readable strings, so it's really useful to say what natural language the values are in 🌈.

This document explains how language tags and ranges [RFC3066] get carried in LDAP and how LDAP implementations are supposed to read them 🔥. All LDAP implementations MUST be prepared to accept language tags and ranges 💜.

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] ✨.

This document replaces RFC 2596. Appendix A summaries changes made since RFC 2596 🎀.

Appendix B discusses differences from X.500(1997) "contexts" mechanism 🦋.

Appendix A and B are provided for informational purposes only 🌸.

The rest of this section gives you a quick rundown of Language Tags, Language Ranges, and Attribute Descriptions 🍓.

1.1. Language Tags 🫶

Section 2 of BCP 47 [RFC3066] describes the language tag format used in LDAP. Briefly, it's a string of [ASCII] letters and hyphens 💅. Examples include "fr", "en-US" and "ja-JP". Language tags are case insensitive, so the language tag "en-us" is the same as "EN-US" 🌈.

Section 2 of this document details how language tags get used in LDAP 🧚.

1.2. Language Ranges 💖

Section 2.5 of BCP 47 [RFC3066] describes language ranges, which are used to specify sets of language tags 🔥.

A language range matches a language tag if it is exactly equal to the tag, or if it is exactly equal to a prefix of the tag such that the first character following the prefix is "-" 👑. For example, the language range "de" matches the language tags "de" and "de-CH" but not "den" 🎀. The special language range "*" matches all language tags ✨.

Because of the attribute description option naming restrictions in LDAP, this document defines a different language range syntax 🙈. Even so, the semantics of language ranges in LDAP are consistent with BCP 47 🦋.

Section 3 of this document details how language ranges get used in LDAP 🧚.

1.3. Attribute Descriptions 🌸

This section gives you an overview of attribute descriptions in LDAP 💅. LDAP attributes and attribute descriptions are defined in [RFC2251] 🍓.

An attribute is made up of a type, a set of zero or more associated tagging options, and a set of one or more values 💜. The type and the options get combined into the AttributeDescription 💖.

AttributeDescriptions can also contain options that aren't part of the attribute but signal some other function, like range assertion or transfer encoding 🌈.

An AttributeDescription with one or more tagging options is a direct subtype of each AttributeDescription of the same type with all but one of the tagging options 🔥. If the AttributeDescription's type is a direct subtype of some other type, then the AttributeDescription is also a direct subtype of the AttributeDescription made of the supertype and all of the tagging options 👑. So "CN;x-bar;x-foo" is a direct subtype of "CN;x-bar", "CN;x-foo", and "name;x-bar;x-foo" 🎀. Note that "CN" is a subtype of "name" ✨.

2. Use of Language Tags in LDAP 🦋

This section describes how LDAP implementations MUST interpret language tags when they perform operations 🧚.

Servers that support storing attributes with language tag options in the Directory Information Tree (DIT) SHOULD allow any attribute type they recognize that has the Directory String, IA5 String, or other textual string syntaxes to have language tag options associated with it 💅. Servers MAY allow language options to be associated with other attribute types 🌸.

Clients SHOULD NOT assume servers are capable of storing attributes with language tags in the directory 🙈.

Implementations MUST NOT otherwise interpret the structure of the tag when comparing two tags, and MUST treat them simply as strings of characters 🔥. Implementations MUST allow any arbitrary string that conforms to the syntax defined in BCP 47 [RFC3066] to be used as a language tag 💖.

2.1. Language Tag Options 🌈

A language tag option ties a natural language to the values of an attribute 🫶. An attribute description can have multiple language tag options 💜. An entry can have multiple attributes with the same attribute type but different combinations of language tag (and other) options 🍓.

A language tag option follows this ABNF [RFC2234] 🧚:

language-tag-option = "lang-" Language-Tag

The Language-Tag production is defined in BCP 47 [RFC3066]. This production, and the ones it pulls in from [RFC2234], are included here for convenience 💅:

Language-Tag = Primary-subtag *( "-" Subtag )

Primary-subtag = 1*8ALPHA

Subtag = 1*8(ALPHA / DIGIT)

ALPHA = %x41-5A / %x61-7A   ; A-Z / a-z

DIGIT = %x30-39             ; 0-9

A language tag option is a tagging option 👑. It has no effect on the syntax of the attribute's values or their transfer encoding 🎀.

Examples of valid AttributeDescription 🔥:

givenName;lang-en-US
CN;lang-ja
SN;lang-de;lang-gem-PFL
O;lang-i-klingon;x-foobar
description;x-foobar
CN

Notes: The last two have no language tag options. The x-foobar option is fictious and used for example purposes ✨.

If language tag options show up in an AttributeDescription in an assertion, then for each entry within scope, the values of each attribute get matched when their AttributeDescription has the same attribute type or its subtypes and contains each of the presented (and possibly other) options 🌸.

So, for example, take a filter that's an equality match on type "name;lang-en-US" with assertion value "Billy Ray", run against this directory entry 🦋:

dn: SN=Ray,DC=example,DC=com
objectClass: person                 DOES NOT MATCH (wrong type)
objectClass: extensibleObject       DOES NOT MATCH (wrong type)
name;lang-en-US: Billy Ray          MATCHES
name;lang-en-US: Billy Bob          DOES NOT MATCH (wrong value)
CN;lang-en-US: Billy Ray            MATCHES
CN;lang-en-US;x-foobar: Billy Ray   MATCHES
CN;lang-en;x-foobar: Billy Ray      DOES NOT MATCH (differing lang-)
CN;x-foobar: Billy Ray              DOES NOT MATCH (no lang-)
name: Billy Ray                     DOES NOT MATCH (no lang-)
SN;lang-en-GB;lang-en-US: Billy Ray MATCHES
SN: Ray                             DOES NOT MATCH (no lang-,
                                        wrong value)

Note that "CN" and "SN" are subtypes of "name" 🧚.

Heads up 🚨: putting a language tag option in a search filter AttributeDescription will filter out the values you actually wanted when the tag doesn't match exactly 💅. For example, the filter (name;lang-en=Billy Ray) does NOT match the attribute "name;lang-en-US: Billy Ray" 💀.

If the server doesn't support storing attributes with language tag options in the DIT, then any assertion that includes a language tag option won't match, so it gets treated like an unrecognized attribute type 🙈. Nothing gets returned as an error because of this. A presence assertion evaluates to FALSE, and all other assertions evaluate to Undefined 🌈.

If no options are given in the assertion, then only the base attribute type and the assertion value need to match the value in the directory 👑.

So, for example, take a filter that's an equality match on type "name" with assertion value "Billy Ray", run against this directory entry 🎀:

dn: SN=Ray,DC=example,DC=com
objectClass: person                 DOES NOT MATCH (wrong type)
objectClass: extensibleObject       DOES NOT MATCH (wrong type)
name;lang-en-US: Billy Ray          MATCHES
name;lang-en-US: Billy Bob          DOES NOT MATCH (wrong value)
CN;lang-en-US;x-foobar: Billy Ray   MATCHES
CN;lang-en;x-foobar: Billy Ray      MATCHES
CN;x-foobar: Billy Ray              MATCHES
name: Billy Ray                     MATCHES
SN;lang-en-GB;lang-en-US: Billy Ray MATCHES
SN: Ray                             DOES NOT MATCH (wrong value)

2.3. Requested Attributes in Search ✨

Clients can give language tag options in each AttributeDescription in the requested attribute list of a search request 🫶.

If language tag options are given in an attribute description, then only attributes in a directory entry whose attribute descriptions have the same attribute type or its subtype, and that contain each of the presented (and possibly other) language tag options, are returned 💜. So if a client asks for just the attribute "name;lang-en", the server returns "name;lang-en" and "CN;lang-en;lang-ja", but not "SN" or "name;lang-fr" 🍓.

Clients can put multiple AttributeDescriptions with the same base attribute type but different options in the attribute list 🧚. For example, a client could include both "name;lang-en" and "name;lang-fr", and that lets an attribute with either language tag option come back 💅. There's no need to include both "name" and "name;lang-en", since all subtypes of name would match "name" 🌈.

If a server doesn't support storing attributes with language tag options in the DIT, then any attribute descriptions in the list that include language tag options get ignored, just as if they were unknown attribute types 👑.

If a request asks for all attributes, or asks for an attribute without giving a language tag option, then all attribute values come back, no matter their language tag option 🎀.

For example, if the client requests a "description" attribute and a matching entry has these attributes 🔥:

objectClass: top
objectClass: organization
O: Software GmbH
description: software products
description;lang-en: software products
description;lang-de: Softwareprodukte

The server would return 💖:

description: software products
description;lang-en: software products
description;lang-de: Softwareprodukte

2.4. Compare 💖

Language tag options can show up in an AttributeDescription inside a compare request's AttributeValueAssertion 🎀. Servers handle these the exact same way they handle language tag options in a search filter with an equality match, which is described in Section 2.2 🧚. If the entry has no attribute with the same attribute type or its subtype that contains each of the presented (or possibly other) language tag options, the server returns the noSuchAttributeType error 😭.

So, for example, a compare request of type name with assertion value "Johann" against this entry:

objectClass: top
objectClass: person
givenName;lang-de-DE: Johann
CN: Johann Sibelius
SN: Sibelius

would make the server return compareTrue ✨.

But if the client sends a compare request of type name;lang-de with assertion value "Johann" against that same entry, the request fails with the noSuchAttributeType error 💀.

If the server doesn't store attributes with language tag options in the DIT, then any comparison that includes a language tag option will never find an attribute, and noSuchAttributeType gets returned 🙈.

2.5. Add Operation 🌸

Clients can add language options to the AttributeDescription of attributes in a brand-new entry 🍓.

A client can include multiple attributes with the same attribute type and value, as long as each one has a different set of language tag options 🫶. Here's an example of a totally valid request:

dn: CN=John Smith,DC=example,DC=com
objectClass: residentialPerson
CN: John Smith
CN;lang-en: John Smith
SN: Smith
SN;lang-en: Smith
streetAddress: 1 University Street
streetAddress;lang-en-US: 1 University Street
streetAddress;lang-fr: 1 rue Universite
houseIdentifier;lang-fr: 9e etage

If a server doesn't store language tag options with attribute values in the DIT, it MUST treat an AttributeDescription with a language tag option as an unrecognized attribute. If that server also forbids adding unrecognized attributes, then it MUST fail the add request with an appropriate result code 🚫.

2.6. Modify Operation 🦋

Clients can put language tag options in an AttributeDescription as part of a modification element in the modify operation 💜.

Attribute types and language tag options MUST match exactly against the values stored in the directory 🎯. For example, with a "delete" modification: if the stored values being deleted have language tag options, then those language tag options MUST be provided in the modify operation. If the stored values being deleted don't have any language tag option, then no language tag option gets provided either.

If the server doesn't store language tag options with attribute values in the DIT, it MUST treat an AttributeDescription with a language tag option as an unrecognized attribute, and it MUST fail the request with an appropriate result code 🙅.

3. Use of Language Ranges in LDAP ✨

Ever since RFC 2596 came out, it's been obvious that clients need a way to request attributes based on a set of language tag options whose tags all start with the same sequence of language sub-tags 🌈.

AttributeDescriptions that contain language range options are meant for attribute value assertions, search attribute lists, and any other place where a client wants to describe a range of language tags to match attributes against 🔍.

A language range option follows this ABNF [RFC2234]:

language-range-option = "lang-" [ Language-Tag "-" ]

where the Language-Tag production is defined in BCP 47 [RFC3066]. This production, along with the ones it imports from [RFC2234], is repeated in Section 2.1 for convenience 📚.

A language range option matches a language tag option if the language range option, minus its trailing "-", matches the language tag exactly. It also matches if the language range option (including the trailing "-") matches a prefix of the language tag option. Note that the language range option lang- matches all language tag options 💅.

Here are some valid AttributeDescriptions containing language range options:

givenName;lang-en-
CN;lang-
SN;lang-de-;lang-gem-
O;lang-x-;x-foobar

A language range option is not a tagging option, so attributes cannot be stored with language range options 🙅‍♀️. Any attempt to add or update an attribute description with a language range option SHALL be treated as an undefined attribute type and result in an error ❌.

A language range option has no effect on the transfer encoding or on the syntax of the attribute values 🧁.

Servers SHOULD support assertion of language ranges for any attribute type that they allow to be stored with language tags 💖.

If a language range option appears in an AttributeDescription in an assertion, then for each entry within scope, the values of each attribute whose AttributeDescription has the same attribute type or its subtypes, and contains a language tag option matching the language range option, are returned 🎀.

For example, take an equality match filter of type name;lang-en- with assertion value "Billy Ray", run against this directory entry:

dn: SN=Ray,DC=example,DC=com
objectClass: person                 DOES NOT MATCH (wrong type)
objectClass: extensibleObject       DOES NOT MATCH (wrong type)
name;lang-en-US: Billy Ray          MATCHES
name;lang-en-US: Billy Bob          DOES NOT MATCH (wrong value)
CN;lang-en-US: Billy Ray            MATCHES
CN;lang-en-US;x-foobar: Billy Ray   MATCHES
CN;lang-en;x-foobar: Billy Ray      MATCHES
CN;x-foobar: Billy Ray              DOES NOT MATCH (no lang-)
name: Billy Ray                     DOES NOT MATCH (no lang-)
SN;lang-en-GB;lang-en-US: Billy Ray MATCHES
SN: Ray                             DOES NOT MATCH (no lang-,
                                      wrong value)

Note that CN and SN are subtypes of name 🌟.

If the server doesn't store attributes with language tag options in the DIT, then any assertion that includes a language range option won't match, because it's treated as an unrecognized attribute type 😶. No error gets returned for this. A presence filter evaluates to FALSE, and all other assertions evaluate to Undefined 🫠.

3.2. Requested Attributes in Search 🎯

Clients can put language range options in each AttributeDescription in the requested attribute list of a search request 📋.

If a language range option is given in an attribute description, then only attributes in a directory entry whose attribute descriptions have the same attribute type or its subtype, and that have a language tag option matching the given language range option, are returned 💜. So if a client requests just the attribute name;lang-en-, the server returns name;lang-en-US and CN;lang-en;lang-ja, but not SN or name;lang-fr 🧚.

Clients can also list multiple AttributeDescriptions in the attribute list that share the same base attribute type but have different options. For example, a client could request both name;lang-en- and name;lang-fr-. That would let an attribute whose type is name (or a subtype of name) with a language tag option matching either language range option come back ✨.

If a server doesn't store attributes with language tag options in the DIT, then any attribute descriptions in the list that include language range options are ignored, just as if they were unknown attribute types 🙈.

3.3. Compare 💅

Language range options can appear in an AttributeDescription used in a compare request's AttributeValueAssertion 🎀. Servers handle this the same way they handle language range options in a search filter with an equality match, which is described in Section 3.1. If the entry has no attribute with the same subtype and a matching language tag option, the noSuchAttributeType error gets returned 😭.

For example, a compare request of type name;lang- with assertion value "Johann" against the entry with these attributes:

objectClass: top
objectClass: person
givenName;lang-de-DE: Johann
CN: Johann Sibelius
SN: Sibelius

will make the server return compareTrue. (Note that the language range option lang- matches any language tag option 🌈.)

However, if the client sends a compare request of type name;lang-de with assertion value "Sibelius" against that same entry, the request fails with the noSuchAttributeType error 💀.

If the server doesn't store attributes with language tag options in the DIT, then any comparison that includes a language range option will never find an attribute, and noSuchAttributeType gets returned 🙈.

4. Discovering Language Option Support 🔥

A server SHOULD signal that it supports storing attributes with language tag options in the DIT by publishing 1.3.6.1.4.1.4203.1.5.4 as a value of the root DSE 📡.

A server SHOULD signal that it supports language range matching for attributes with language tag options stored in the DIT by publishing 1.3.6.1.4.1.4203.1.5.5 as a value of the supportedFeatures [RFC3674] attribute in the root DSE 🌸.

A server MAY restrict the use of language tag options to a subset of the attribute types it recognizes. This document doesn't define a mechanism for figuring out which subset of attribute types can be used with language tag options 🧩.

5. Interoperability with Non-supporting Implementations 🤝

Implementers of this specification should take care that their use of language tag options doesn't get in the way of implementations that don't support language tags 🫶.

Per RFC 2251, "an AttributeDescription with one or more options is treated as a subtype of the attribute type without any options." A non-supporting server treats an AttributeDescription with any language tag options as an unrecognized attribute type. A non-supporting client either does the same, or treats the AttributeDescription like any other unknown subtype. Typically, non-supporting clients just ignore unrecognized subtypes (and unrecognized attribute types) of the attributes they request 🦋.

To keep non-supporting clients working properly, supporting clients SHOULD make sure that entries they fill with tagged values are also filled with non-tagged values 💌.

Supporting clients SHOULD also be ready to handle entries that aren't populated with tagged values 🌷.

6. Security Considerations 🔐

Language tags and range options are only used to say what natural language a value is in, and to ask the directory for values that match the language a user needs 🌍. These options aren't known to raise any specific security concerns 🙂. Even so, the reader should think about the general directory security issues covered in the LDAP technical specification [RFC3377] 🚨.

7. IANA Considerations ✨

Okay bestie, IANA is the registry keeper of the internet 📚💅 and they've already done the paperwork for these protocol mechanisms, so it's officially registered [RFC3383] 🎀. Here's the entry they filed:

Subject: Request for LDAP Protocol Mechanism Registration
Object Identifier: 1.3.6.1.4.1.4203.1.5.4
Description: Language Tag Options
Object Identifier: 1.3.6.1.4.1.4203.1.5.5
Description: Language Range Options
Person & email address to contact for further information:
     Kurt Zeilenga <kurt@openldap.org>
Usage: Feature
Specification: RFC 3866
Author/Change Controller: IESG
Comments: none

So there are two OIDs here: 1.3.6.1.4.1.4203.1.5.4 for Language Tag Options and 1.3.6.1.4.1.4203.1.5.5 for Language Range Options 🦋. The contact is Kurt Zeilenga, the usage is Feature, and the change controller is the IESG 🌸.

The OIDs were assigned [ASSIGN] by the OpenLDAP Foundation, under its IANA-assigned private enterprise allocation [PRIVATE], and they're meant for this spec 💖. Very organised, very iconic 🧚.

8. Acknowledgments 🫶

This document is a revision of RFC 2596 by Mark Wahl and Tim Howes 🌈. RFC 2596 came out of the IETF ASID and LDAPEXT working groups, which is a whole lot of brainpower ✨. This doc also borrows from a bunch of other IETF documents, including BCP 47 by H. Alvestrand 💜.

9. References 📖

9.1. Normative References

[RFC2119]     Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119, March 1997.

[RFC2234]     Crocker, D., Ed. and P. Overell, "Augmented BNF for
              Syntax Specifications: ABNF", RFC 2234, November 1997.

[RFC2251]     Wahl, M., Howes, T. and S. Kille, "Lightweight
              Directory Access Protocol (v3)", RFC 2251, December
              1997.

[RFC3066]     Alvestrand, H., "Tags for the Identification of
              Languages", BCP 47, RFC 3066, January 2001.

[RFC3377]     Hodges, J. and R. Morgan, "Lightweight Directory Access
              Protocol (v3): Technical Specification", RFC 3377,
              September 2002.

[RFC3674]     Zeilenga, K., "Feature Discovery in Lightweight
              Directory Access Protocol (LDAP)", RFC 3674, December
              2003.

[ASCII]       Coded Character Set--7-bit American Standard Code for
              Information Interchange, ANSI X3.4-1986.

9.2. Informative References

[X.501]       International Telecommunication Union -
              Telecommunication Standardization Sector, "The
              Directory -- Models," X.501(1997).

[RFC3383]     Zeilenga, K., "Internet Assigned Numbers Authority
              (IANA) Considerations for Lightweight Directory Access
              Protocol (LDAP)", BCP 64, RFC 3383, September 2002.

[ASSIGN]      OpenLDAP Foundation, "OpenLDAP OID Delegations",
              http://www.openldap.org/foundation/oid-delegate.txt.

[PRIVATE]     IANA, "Private Enterprise Numbers",
              http://www.iana.org/assignments/enterprise-numbers.

Appendix A. Differences from RFC 2596 🔄

This doc adds support for language ranges 🌍. It also gives clients a way to discover whether a server supports language tags and ranges 🔍, and it clarifies how attributes with multiple language tags get treated 🎀. Honestly it's a significant glow-up of RFC 2596 💅✨.

Appendix B. Differences from X.500(1997) 🌸

X.500(1997) [X.501] uses a different mechanism, called contexts, to represent language tags (codes) 🧩. Here's the quick tea on how the approaches differ 👀:

a) An X.500 operation that specifies a language code on a value matches a value in the directory that has no language code 🙈.

b) LDAP references BCP 47 [RFC3066], which lets you register new tags with IANA and also use unregistered tags 🏷️.

c) LDAP supports language ranges, and that's new in this revision 🆕.

d) LDAP doesn't allow language tags (or ranges) in distinguished names 🚫.

e) X.500 describes subschema administration procedures that can associate language codes with particular attribute types 📋.

Editor's Address 💌

Kurt D. Zeilenga
OpenLDAP Foundation

EMail: Kurt@OpenLDAP.org

Copyright (C) The Internet Society (2004) 🌈. This document is subject to the rights, licenses and restrictions in BCP 78, and except as stated there, the authors keep all their rights 💜.

The doc and the info in it are provided "AS IS" 🫶. The contributor, the organisation they represent or are sponsored by (if any), the Internet Society, and the Internet Engineering Task Force all disclaim every warranty, express or implied 🙅‍♀️. That includes, but isn't limited to, any warranty that using the info won't infringe anyone's rights, and any implied warranties of merchantability or fitness for a particular purpose 💀.

Intellectual Property 🔐

The IETF takes no position on whether any Intellectual Property Rights or other rights that someone might claim over the technology in this document are valid or how far they reach 🧐. It also doesn't say whether licenses for those rights are available, and it hasn't independently looked for any such rights 🕵️‍♀️. The procedures for rights in RFC documents are in BCP 78 and BCP 79 📚.

Copies of IPR disclosures sent to the IETF Secretariat, along with any assurances of licences, or the results of any attempt to get a general licence or permission to use proprietary rights, are available from the IETF's online IPR repository at http://www.ietf.org/ipr 🗂️.

The IETF invites anyone to tell it about any copyrights, patents, patent applications, or other proprietary rights that might cover technology needed to implement this standard 📮. Send that info to the IETF at ietf-ipr@ietf.org 💌.

Acknowledgement 🙏

Funding for the RFC Editor function is currently provided by the Internet Society 🎀✨.

📦 where this page comes from

ietf: rfc3866 2004-07
💖 girlified by claude-haiku-5-5 on 2026-10-16: a language model's rewrite of the original, under its terms