Dear colleagues in CGP/JGP/KGP, Below I send you my homework (action item 2) which is "writing on the number of allocatable variants" issue. I divided it into 2 sub-issues as below. Hiro ============= issues regarding "the number of allocatable variants" ============= proposal by IP: the number of allocatable variant strings should be small, e.g., at most 3, as the result of RootLGR issue-1 "what kinds of variant strings are preferred to be applied for" cannot be predicted and should not be assumed in RootLGR (at least in Japanese case, for example) <reasoning> In Japanese case, all the characters are considered to be independent. Namely, all the strings of any combination of any characters in the repertoire are designed to be allocatable by different TLD applicants. However, JGP knows that CGP wants to define variants in RootLGR, and JGP currently intends to compromise by basically importing variant definitions from CGP. This means that such strings that have variants in C will be packaged also in J, although they are originally intended to be independent. From Japanese point of view, as all the characters are originally independent, all the strings in the package should be allocatable even if they can be delegated to the same applicant. Moreover, the probability of some applicant's intention to apply for a specific string in the package is equal for all the strings in the package. issue-2 If limitation of the number of delegated variant strings is intended to be implemented, it should be done in application phase or application evaluation phase - not by RootLGR <reasoning> As discussed in (1), the probability of "which string is preferred" depends on the specific string that is applied for or that is accessed by the users. Namely, it depends on how much the specific string is well-known to the users. This means it's not possible for us to define WLE before asking each applicant about which variant strings they want to apply for. In that case, if we want to limit the number of delegated strings or to assess whether it is appropriate for the applied-for string picked up from the package, we need to engage measures after application(s) of specific string(s) is(are) made. ==