Policy for IDN Code Points

The VeriSign Shared Registration System (SRS) allows a registrant to register IDNs through a registrar in any script supported by Unicode. To allow for rare scripts, musical notations, and other special characters, VeriSign has developed a Policy for IDN Code Points specifying permissible, restricted, and prohibited code points.

Part I: Permissible Code Points

The VeriSign implementation of IDN registration for the .com, .net, .name, .tv, and .cc top-level domains (TLDs) follows the guidelines set forth by the Internet Engineering Task Force (IETF) in these RFC documents:

• RFC 3490 -- IDNs in Applications (IDNA)
• RFC 3491 -- Name Preparation (nameprep)
• RFC 3454 -- String Preparation (stringprep)
• RFC 3492 -- Encoding Scheme (punycode)

The VeriSign registries for .com, .net, .name, .tv, and .cc allow registration of code points included in Unicode 3.2 subject to the restrictions in the above RFC documents.

Part II: Restricted Code Points

The .com, .net, .name, .tv, and .cc registries implement certain restrictions on code points beyond that allowed by the RFC documents. Code points introduced after Unicode 3.2 are prohibited. IDN objects must be properly encoded in Punycode and start with the “xn--“ prefix as identifier indicating they are IDNs. Domain names that start with 2 ASCII letters immediately followed by 2 consecutive hyphens other than “xn--“ are not allowed for registrations.

Part III: Further Restrictions on Commingling of Scripts

As a rule, the .com, .net, .name, .tv, and .cc registries reject the commingling of code points from different Unicode blocks. That is, if an IDN contains code points from any two or more Unicode blocks, then that IDN registration is rejected. For example, letter(s) from Latin alphabets cannot be used with letter(s) from Cyrillic in an IDN. All code points within an IDN must come from the same single Unicode block.

An exception to this rule is made for well-defined languages which want to allow this commingling in certain circumstances. For instance, symbols for the Korean language are contained in the Unicode block called “Hangul Syllables” (u+AC00 – u+D7AF), however, Korean commonly uses basic Latin characters as well. National Internet Development Agency (NIDA) of Korea, the .kr ccTLD registry, developed a language table which includes the “Hangul Syllables” as well as lowercase Latin letters (a-z), numbers (0-9) and the hyphen character. Normally these characters would not be allowed to commingle in a single domain name, but IDN objects marked with the “KOR” language tag for Korean can safely contain the code points above.

IDN registrations marked with one of the following language tags will be checked against that language’s definition to determine validity. IDN registrations marked with a language not in the table must meet the zero-commingling conditions of the original rule.