For each Unicode version starting with Unicode 5.1, a table is provided in the versioned directories under [IDNA-Table]. Each table in a version of the Unicode standard is always backwards compatible with previous versions of the table: only characters whose state value is not allowed can change state or mapping value. Unlike the IDNA2008 table, this table is designed to apply to the entire domain name, not just individual labels. This design allows IDNA2003 management of label separators. In particular, the table is structured in such a way that problematic characters such as U+2488 ( 1. ) DIGIT ONE POINT whose decompositions contain a “point” are prohibited. The specification offers two main features: one is a comprehensive mapping to meet users` current expectations for capitalization and other domain name variants. Such mapping is made possible by IDNA2008. The second is a compatibility mechanism that supports existing domain names authorized by IDNA2003. This second feature is designed to improve customer behavior during the transition period. The IDNA mechanism for allowing non-ASCII Unicode characters in domain names involves applying the following steps to each label in the domain name that contains Unicode characters: UTS #46 defines a mapping that meets the normative requirements of the IDNA2008 protocol and is as compatible as possible with IDNA2003. For client software, this provides behavior that best matches user expectations for processing domain names with existing data, that is, domain names are mapped consistently across clients that support IDNA2003 and clients that support IDNA2008 with UTS mapping #46.

The term “registers” encompasses much more than top-level registers, such as: for .de or .com. For example, .blogspot.com has registered more domain names than most top-level registries. Different policies may apply to a record and its sub-registrations. Therefore, millions of registries must be included in a transition strategy, not just hundreds. Due to the visual confusion introduced by joiner characters, IDNA2008 provides them with a special category called CONTEXTJ and only allows CONTEXTJ characters in limited contexts: certain sequences of Arabic or Indian characters. However, applications that run IDNA2008 lookup do not need to search for these contexts, so overall security depends on correct registry implementations. In addition, IDNA2008 context constraints do not capture most cases where different domain names appear visually confusing due to ZWJ and ZWNJ. The input for Unicode IDNA compatibility processing is a prospective domain_name string expressed in Unicode and a choice between transition and non-transition processing. The domain name consists of a sequence of labels delimited by periods, such as “Bücher.de”. For more information on composing a URL, see Section 3.5 of [STD13]. UTS #46 provides a compatibility bridge that allows implementations to manage IDNA2003 and IDNA2008 domain names. Specifications and other general information can be found at UTS #46.

Any compliant implementation may also have stricter validity criteria than those in Section 4.1, Validity Criteria. For example, an application might prohibit or warn against domain name labels with certain characteristics, for example: B: With the exception of line (g), characters that are mapped/ignored in a specification (lines (d), (e), (k)) are either mapped/ignored in exactly the same way, or not allowed in the other specifications. This prevents domain names from being assigned differently in different browsers: either the characters are assigned to the same result or they do not work. Line (k) is not problematic in this respect, provided that the registers follow one of the specifications, since characters such as U+023A ( ⱥ ) are not valid on the registered labels. Domain names such as “macchiati.blogspot.com” were originally developed to support only ASCII characters. In 2003, the first specification was published that allows the use of most Unicode characters in domain names. This specification was replaced a few years later by IDNA2008, which differs in some respects. IDNs are supported by all modern browsers and email programs, allowing users to use links in their native language such as Bücher.de. In this document, a label is a substring of a domain name.

This substring is delimited on both sides by the beginning or end of the string, or by one of the following characters, called label delimiters: Any domain_name input string that does not log an error has been successfully processed according to this specification. Conversely, if an input string domain_name causes an error, processing of the input string fails domain_name. The decision on what to do with the error entry rests with the appellant and is outside the scope of this document. The processing is idempotent: reapplying the treatment at the output does not make any additional changes. For examples, see Table 2, Examples of Transitional Treatment. The test format and file name were changed in version 11.0 so that users can express a variety of different combinations of input options. The new format allows the test implementation to accurately test the results of its supported flag combination by filtering status codes that correspond to an unsupported input flag. The XV8 value was also deleted as it was not very useful in practice. IDNA2003 provides a flag, UseSTD3ASCIIRules, that allows implementations to choose whether or not to follow the rules of [STD3]. These rules exclude ASCII characters outside the set of A-Z, A-Z, 0-9, and U+002D ( – ) MINUS HYPHEN. For example, some browsers also allow characters such as U+005F ( _ ) LOW LINE (underbar) in domain names and therefore use UseSTD3ASCIIRules=false as well as their own validation checks for other ASCII characters. IDNA2003 applications typically display the processed string to the user.

This improves safety by reducing the risk of visual confusion. For example, the URL googIe.com (with a capital I instead of the L) is displayed as a googie.com. In case of discrepancies, however, the distinction between the original form and the treated form is particularly important for users. Therefore, non-transient processing should be used to display domain names. This is the same as transition processing, except that deviations: ß and ς and joins are excluded. It is therefore fully compatible with IDNA2008 for these deviation characters. One of the great strengths of domain names is universality. The URL allows Apple.com to access Apple`s website from anywhere in the world and with any browser. The markdavis@google.com email address can be used to send emails to a publisher of this specification from anywhere in the world using any email provider. This compatibility treatment maps a domain name such as ÖBB.at to the valid domain name öbb.at, which meets users` expectations for case-sensitive behavior in domain names.

For transient use, compatibility processing also allows domain names that contain symbols and punctuation marks that were valid in IDNA2003, such as √.com (with which a Web page is associated). These domain names containing symbols will gradually disappear as registrars move to IDNA2008. The processing in IDNA2008 is identical to IDNA2003 for many common domain names. IDNA2003 and IDNA2008 transform a Unicode domain name in an IRI (such as öbb.at) into a Punycode version (such as xn--bb-eka.at). However, IDNA2008 does not maintain strict backward compatibility with IDNA2003. The main differences are: UTS #46: Unicode IDNA Compatibility Processing, sometimes referred to as “TR46”, is a Unicode specification that enables implementations to manage domain names in a compatible manner during the transition from IDNA2003 to IDNA2008. The title is Unicode IDNA Compatibility Processing.