Dispute about new standard for non-English domains

The Internet Engineering Task Force (IETF) is arguing about a successor for the IDNA 2003 standard for non-English domains adopted in 2003. The initiators of a new version pointed out considerable weaknesses of the existing standard and said that adequate representation of some non-English character sets is impossible in the current Domain Name System (DNS). For some sets, both individual characters and the special characters required for their correct representation are missing. However, registry members fear that the planned proposal will cause even more confusion. They are especially concerned because the new proposal makes a number of existing conventions invalid.

Advertisement

Even German users, who are not generally affected by the internationalisation debate, could be confronted with a problem. The letter "ß", which is not included in the current IDNA standard and is currently represented as "ss", is on the list of possible additions to the character set. People who have relied on "ß" being converted into "ss" during a DNS query and have shared their contact details with customers, colleagues or friends accordingly will have to consider that "ß" will have a life of its own, warns Marcos Sanz of German DENIC eG.

Sanz believes that there is not enough evidence to justify a new standard yet. His colleague Stephane Bortzmeyer of French AfNIC agrees. Bortzmeyer was particularly concerned about the new proposal's much stricter implementation guidelines for registries. The representatives of the affected organisations fear that using the standard itself to regulate the registry options available at the time of registration could make registries highly inflexible. Some of them want to allow mapping older conventions on newer ones, for example. They say it would also be virtually impossible to regulate registrations on the 3rd or 4th level with these measures.

Co-author John Klensin and IDN experts Harald Alvestrand and Patrik Fältström propose a stricter classification of acceptable and excluded characters for the new standard. While the acceptable characters need to be part of the Unicode standard as before, the proposed IDNA standard is no longer restricted to one Unicode version - this restriction is possibly at the root of most of the current standard's problems. In addition, IDN supporters want to introduce eight additional character classes.

Some of the proposed characters in these classes correspond to sequencing characters like dots or dashes in the Latin character set. Other special characters like musical notation symbols are to be excluded. There is a category for characters requiring special treatment as well as an empty class incorporated for future use. Standard proposers have even dared to discuss individual characters. However, nobody expects the proposal to result in a perfect representation of all the languages on the net. When the proposal was presented in Philadelphia, a Hebrew language specialist already pointed out that certain restrictions affecting number and special character combinations in domain names written from right to left exclude several Hebrew words.

Developing the standard will be a lot of work, said Ram Mohan, CTO of Afilias, one of the largest registries. While Mohan urgently advised to step away from a fixed Unicode version, he spoke out against changing the xn-- label for Punycode (ASCII) representations of non-Latin domains. At the meeting Mohan warned, "This won't be as easy as some people say". (Monika Ermert) /

(trk)