Topics: 2017
JPRS Submitted Opinion to ICANN on Reference Japanese Label Generation Rules (LGR) for the Second Level
On 16 June 2017, JPRS submitted the letter to ICANN President and CEO Goran Marby, regarding Reference Japanese Label Generation Rules (LGR) for the Second Level that ICANN has been and is facilitating the development.
The submitted letter is the response to the Staff Report[*1] posted on 28 April 2017 that ICANN staff wrote up based on the submitted comments on Public Comment[*2] period.
JPRS submitted the letter mainly focused on the "visual similarity" issue, in their capacity as registry that deals with Japanese IDNs, as part of the cooperative work between ICANN and Japanese community.
■References:
- *1. Staff Report: Reference Japanese Label Generation Rules (LGR) for the Second Level
https://www.icann.org/en/system/files/files/report-comments-japanese-lgr-second-level-28apr17-en.pdf - *2. Public Comment:Reference Japanese Label Generation Rules (LGR) for the Second Level
https://www.icann.org/public-comments/japanese-lgr-second-level-2017-01-27-en
■Overview:
- Summary of the Staff Report
- JPRS's input to the Staff Report
- There are many other similar characters[*] in Japanese, in addition to U+30FC (ー), U+4E00 (一) and U+30FD (ヽ), U+4E36 (丶) which have been defined as variant code points in the proposed reference LGR.
* カ / 力, オ / 才, ロ / 口, ハ / 八, ト / 卜, ニ / 二 or エ, etc.
- The description section of the reference LGR can be updated to explicitly include those similar characters.
- Though this explicit mention in the description section will not be a formal variant relation, it will allow the registries to note these and take the appropriate action of their choice to mitigate any impact on users.
- This change will be discussed with the stakeholders, especially with the Japanese language community for finalization.
- Visual similarity exists in other languages[*]. It is not the matter of specific languages.
* 1 / l, 0 / O, etc.
- All the allocatable characters and labels are regarded as independent and no variants are defined in Japanese IDNs under .jp for more than 15 years.
- The selection of characters and labels can be respected without disrupting the security and stability of DNS.
- We oppose to introduce visual similarity as variant into the reference LGR even in the description section.
- If coping with visual similarity is needed, it should be controlled in the domain name usage sphere and be handled by applications or DRP, not by blocking domain name registration.
Letter
16 June 2017
Goran Marby Dear ICANN President and CEO Goran Marby, We appreciate the efforts ICANN has been and is facilitating the development of Reference LGRs for the second level. Please receive input from a registry that deals with Japanese IDNs, as part of the cooperative work between ICANN and Japanese community mentioned in the last paragraph of the staff report[1] posted on April 28 2017. [1]https://www.icann.org/en/system/files/files/report-comments-japanese-lgr-second-level-28apr17-en.pdf First, please note the background why we did not post our comments to the second public comment invitation[2]. We considered the then-proposed LGR had some issues from the viewpoint of Japanese community but was acceptable as a reference LGR to gTLDs in general, because reference is a reference, and sensible deviation from it is allowed. For example, some modern notations are not allowed by the proposed LGR, but they will be allowed in some TLDs. [2]https://www.icann.org/news/announcement-2017-01-27-en Secondly, we'd like to comment on introducing visual similarity into domain registration rules. The Google Registry states that the limitation of registration of labels with visually similar Japanese characters may reduce the risk of user confusion. If ICANN seeks to deal with visual similarity at the level of domain name registration rules, it is needed to (re)visit LGRs of all the languages to see whether visual similarity is well-considered. That is, visual similarity is not the matter of specific languages. In addition, there is no well-established definition of visual similarity, including homoglyphs. We think it is disastrous to all the whole LGR space if ICANN introduces the concept of visual similarity to the LGR. Thirdly, let us refer to the substance of Google Registry's comments. Japanese community has been using Japanese IDNs under .jp for more than 15 years. Over 110 thousand Japanese IDNs are currently registered and used. The rules of Japanese IDNs were developed based on the consultation with Japanese community. All the allocatable characters and labels are regarded as independent and no variants are defined. That is, no matter how label1 and label2 are visually similar, the domain name registrant of each label has equal rights about the domain name and registrant of label1 does not have the right to block the registration and usage of label2. 15+ years' experience has shown us that each character and each label in Japanese repertoire has its firm identity and Japanese users can manage the situation well. Based on the above background, we believe that the wishes and intentions of the registrants regarding the selection of characters and labels can be respected without disrupting the security and stability of DNS. Therefore we opine that ICANN should not recommend registries to block domain name registration of label2 on the grounds that label2 looks visually similar to label1 that has already registered as a domain name. In order not to be recognized that domain labels with any visual similarity should be avoided, we oppose to introduce visual similarity as variant into the reference LGR even in the description section. If coping with visual similarity is needed, it should be controlled in the domain name usage sphere and be handled by applications or DRP, not by blocking domain name registration. Sincerely,
Koki Higashida, |