4

The search function finds both voiced and unvoiced kana with either input.

enter image description here

Since these are better considered completely different letters, I am assuming this is a bug.

(Maybe this was implemented to allow for sound changes via rendaku...? But then in English, we don't treat L & N the same, just because in- becomes il- in illegible.)

| |
  • 2
    I'm guessing this might be because the search uses NFD form (and therefore treats the dakuten as a separate character), I went into this at the very end of this answer – cypher Jul 25 '16 at 2:12
1

While I agree with this statement, this may be by design. For some databases, Microsoft SQL Server does this, the dakuten is treated as an "accent" and with certain settings with an accent and without accent are treated as the same character. Note, that this is the same functionality that allows you to find both full-width and half-width kana and ゃ and や with just a single search term.

| |

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .