Linguistic grouping, non latin characters

[simoncn]

Thanks very much for these pointers. As I understand it, there would be 48 additional alpha groups indexed by Gojūon (Hiragana) characters and each group would contain titles starting with the Hiragana character in the group index and also titles starting with the corresponding Katakana character. For example, the group for あ would include titles starting with あ and titles starting with ア. Is this correct?

I have not yet fully understood variants but I think this would add a small number of additional starting characters into some of the 48 alpha groups. For example. the group for つ would include titles starting with っ and ッ. Is this correct?

It looks like it will not be simple for me to implement a solution based on the above proposal that fully meets the requirements for Japanese users. In addition to the requirement to extend the current support for alpha grouping as described above, I am not sure whether the current MinimServer collation sequence (sort order) for Hiragana and Katakana characters is what a Japanese user would expect. At present, the collation sequence for these characters is based on their Unicode character values for performance reasons.

For these reasons, I am wondering whether implementing your alternate proposal of enabling the user to specify a list of additional starting characters for grouping might be a more practical first step. This also has the advantage of potentially being useful for users with titles in other languages that use non-Latin characters.