Serch function doesn't work properly for cards with Kanji name and ruby(syllabic characters)
For example when we search "時を解す者、テフェリー"(Teferi, Time Raveler) at deck building in Japanese version, we couldn't find it with the card name.
We have to input additional characters(ruby) to find the card with the current searching system.
[時を解す者、テフェリー] is like that [時(とき)を解(ほぐ)す者(もの)"、"テフェリー].
It's too difficult to search.
We want to find out cards with card names themselves with Kanji and/or its card name pronunciation without Kanji.
ziguyan#66212
shared this idea
4 comments
-
Jivan#08017
commented
This comment is aimed directly at the game devs.
The problems with sorting/searching Japanese strings are well-documented, and (to my knowledge) Unicode does not describe anything that attempts to tackle this problem. IMO, the best solution that can be easily implemented would be to store the following extra fields in the database:
(a) "ruby-less" name: card name without ruby/furigana, e.g. 「時を解す者、テフェリー」; and
(b) "kanji-less" name: card name with kanji replaced by their hiragana pronunciation, e.g. 「ときをほぐすもの、テフェリー」.
This allows users to either:
(a) type a substring of the ruby-less name if they know all of the kanji in that substring; or
(b) type a substring of the kanji-less name if they don't know all of the kanji in that substring, but can recall the pronunciation of that part of the card name. (I think this is particularly useful from a UX standpoint, as a user using Japanese IME input may just start typing the beginning of the card name in romaji, whence hiragana appears in the input field before they would usually press SPACEBAR to convert the typed hiragana characters into mixed text, and thus the desired search result may appear before SPACEBAR is even pressed.)
One caveat of this proposed solution is that it does not allow a user to input a substring of the name where they use kanji in some places, but the furigana pronunciation in others. For example, 「時をほぐ」 would match neither the ruby-less nor kanji-less name; in this case, the user does not know 「ほぐ」=「解」, but it can be expected that they know 「時」=「とき」, and thus they can be expected to type 「ときをほぐ」 instead to get the result they are searching for. IMO, this is an acceptable caveat.
Implementation-wise, such a solution should be simple: assuming a SQL database, simply add SQL views to the database or generated columns to the existing table of cards, which compute the kanji-less and ruby-less text from the overloaded text (that is, from the currently stored card names, e.g. 「時(とき)を解(ほぐ)す者(もの)、テフェリー」), as follows:
(a) compute the ruby-less text from the overloaded text by removing all parenthesised substrings; and
(b) compute the kanji-less text from the overloaded text by removing all characters that are a parenthesis or a kanji — Unicode provides algorithms for identifying kanji characters, so your programming language should already have a function that can do this for you.
-
DthGun#84182
commented
This bug shold be fixed immediately.
I think people who can't use japanese can't understand what the idia says.This is so serious problem.
Example, you want to search [ Teferi, Hero of Dominaria ].
So you searched with [ Teferi ]/[Hero]/[Dominaria].
But nothing matches.If you want to search [ Teferi, Hero of Dominaria ],
You have to enter [tefe(ə)ri, He(hi(ə))ro(rou) of(əv) Do(dou)mi(mí)na(nə)ri(rí)a(:ə)What do you think?
I think this is garbage search engine. -
Laguz#66798
commented
When I use Unmoored Ego or Sorcerous Spyglass, I have often taken my times in MtGA.
-
lighdar#58254
commented
This is really stupid.
please imagine you should type "cat /kæt/" instead of "cat", every time you search them.
this is what Japanese user should do now.