]> Keio University

Graduate School of Media Design 4-1-1 Hiyoshi, Kohoku-ku Yokohama Kanagawa 223-8526 Japan +81 45 564 2517 t.nemo10@kmd.keio.ac.jp

JPRS

Chiyoda First Bldg. East 13F 3-8-1 Nishi-Kanda Chiyoda-ku Tokyo 101-0065 Japan +81 3 5215 8451 yoshiro.yoneya@jprs.co.jp

This document reports implementation experience of IDNA2008 and findings from the implementation.

This document is a report of implementation experience to evaluate whether IDNA2008 can implement, IDNA2008 implementation have Unicode agility, and providing mapping methods based on and for applications are useful. In this document, two implementations based on IDNA2008 are described. The one provides support functions for various manipulations of internationalized domain names. It is to evaluate whether core protocol of IDNA2008 can implement and the implementation provides useful mappings for applications. The other generates idnabis-tables.xml by using Unicode properties files. It is to evaluate whether IDNA2008 implementation has Unicode agility. This document reports implementation experience of IDNA2008 and findings from the experience.

Purposes of the implementations are following; To evaluate whether core protocol of IDNA2008 can implement. To evaluate whether IDNA2008 implementation has Unicode agility. To evaluate whether providing mappings of and for applications are useful.

In this document, two implementations based on IDNA2008 are described. The one based on core protocol of IDNA2008 provides APIs for handling internationalized domain names. Those APIs are for A/U-label conversion, two labels comparison and label validity check. The implementation is to evaluate whether core protocol of IDNA2008 can be implemented and providing mappings for applications are useful. The implementation also provide three command for convert, compare and check of internationalized domain names for end users. The implementation can import tables of certain version of Unicode properties and corresponding idnabis-tables.xml for codepoint validation. The other implementation generates idnabis-tables.xml based on by using Unicode properties files. The idnabis-tables.xml's format is the same as idnabis-tables.xml published by IANA. This implementation is for the APIs can have Unicode agility by using idnabis-tables. The generator is for evaluating latest Unicode version before IANA publish idnabis-tables.xml for the Unicode version.

APIs |+--------+ +--------+ +--------+ | | | -convert |+--------+ | | | | | | | (A <--> U) || | | Wrapper| | XS | | | | -compare || JNA | | | | | | | | (A/U <--> A/U) |+--------+ +--------+ +--------+ | | | -check |+--------------------------------+ | | (A/U --> Valid/Invalid) || | | || idnkit / idnkitlite (C/C++ API) | | +------------------------+ || <------|Unicode properties files| |+----------------------------------------+ | |IANA idnabis-tables.xml | +-------------------------------------------+ +------------------------+ +----------------------------------------+ import files | OS | +----------------------------------------+ ]]> fig. components of the implementation

Findings from implementation experiences are following. IDNA2008 core protocol can implement. IDNA2008 implementation can have Unicode agility. Originally, it complied with Unicode 5.2.0 based on . And it is easily complying with Unicode 6.0.0 based on . Further, it can comply with Unicode 6.1.0 by using the idnabis-tables.xml generator. This implementation is able to choice a method of mappings based on or . It is expected to improve the usability of user input. And also it can facilitate the transition from IDNA2003 to IDNA2008. But which is preferable mappings of or is not determined yet.

Unicode agility depends on the agility that IANA publishes the latest idnabis-tables.xml when Unicode version was updated. IANA is expected to publish the latest idnabis-tables.xml, when latest Unicode version is published.

The implementation does not warn and handle strings with care when input strings are consisted from mixed scripts. To avoid user's confusion and unwanted access to phishing sites, the application policy about user interface and the domain name registration policy should take into consideration for using mixed scripts.

Patrik Faltstrom provided a program to calculate Unicode codepoints validity according to and generate tables.

&rfc5890; &rfc5891; &rfc5892; &rfc5893; &rfc5894; &rfc5895; &rfc6452;

Updated submission date to avoid expiration.