Summary

  • RZ-LGR conformance is a technical label test, not a finding that a string is a meaningful word.
  • The RZ-LGR need not support every natural-language expression, and mixed-script code points may make a string invalid.

The 2026 Applicant Guidebook explains the boundary directly: DNS labels are identifiers, not a medium for writing a language or its literature. The RZ-LGR is therefore not expected to permit the full expressive range of any natural language, and a string it generates is not required to be a word in a language.

For the 2026 Round, version 6 of the RZ-LGR applies. It contains a distinct label generation rule set for each script or writing system it covers.

Technical conformance remains mandatory. An applied-for IDN must comply with IDNA2008 and the applicable RZ-LGR. A primary string is tested for validity as a TLD, while the RZ-LGR is the sole source for calculating variant strings and their allocatable or blocked dispositions.

This produces two non-equivalences. A label that has clear meaning to a language community does not pass merely because it is meaningful. Conversely, an output permitted by the RZ-LGR need not form a dictionary word.

Composition also matters. The Guidebook says strings mixing code points from LGRs for different scripts may be marked invalid. If the algorithmic check deems a non-conforming string invalid or blocked, the application submission system will not accept it.

BTW recommends separate reviews for language meaning, community recognition and technical label validity, with the applicable RZ-LGR result recorded independently. This is editorial guidance, not an ICANN requirement.

Sources