Why does Calibre treat Japanese as Chinese in filenames?

[Haidon]

The title says it all: Why does Calibre treat Japanese as Chinese in filenames?

This is really annoying. If the program is going to insist on romanizing authors names and titles when it builds its folder hierarchy, it could at least do a proper job of it. I can only assume that the software is treating it all as Chinese rendered in pinyin without any diacritics.

For example, instead of "氷菓 - 米澤 穂信", or if romanization is for some reason necessary (I would have thought that modern computer systems would be able to deal with unicode) at least "Hyouka - Yonezawa Honobu", I end up for example with such filenames as "Bing Guo - Mi Ze Sui Xin" — which to me, someone who cannot read Chinese or pinyin, is indecipherable. I realise that automated romanization of Japanese can be difficult given that kanji can have different readings, but again, why not just use the original data in unicode? As it stands, this behaviour makes no sense whatsoever, and makes it a real downright pain when I want to organise my files manually after copying them to my reader.

Is there some toggle in the program's myriad options to fix this issue?

[kovidgoyal]

Change your user interface language to japanese.

[samhy]

Thank you Haidon for the question (I was wondering the same thing) and kovidgoyal for the answer.

[Haidon]

Originally Posted by kovidgoyal

Change your user interface language to japanese.

Well, it seems that it will now romanize things correctly on import, but I'm afraid that this isn't a very elegant solution. It doesn't fix any previously imported files, and I am sure that there are multilingual users out there who might wish to use the interface in a language other than Japanese. And what if a user were to import a Chinese language book with the interface set to Japanese? I imagine that very similar problems would ensue. I don't see why the software should treat all CJK texts as Chinese unless the software is explicitly set to Japanese. This kind of behaviour makes about as much sense to me as running all English language metadata through a cipher and then encoding it in Cyrillic. It would be nice to see this corrected. Why not just use the unicode data? The software wouldn't even require any romanization algorithm if it would do that. At the very least, most books include a language tag in their metadata –*if for some reason romanization is necessary, why can't Calibre just use that to determine which language it should use to romanize the file data?

[kovidgoyal]

Relying on filenames for metadata is an even less elegant solution, dont do it.

[ctop]

This is a 12 year old post, but the problem still seems to be essentially the same. I have both Chinese and Japanese texts in my library and would like to have them transliterated correctly.
On importing a book, I think Calibre needs to read the file metadata to correctly determine author and title and put it in the right place. At this point, it could also read the language metadata and insert a step of transliteratiion *based on the file language* rather than forcing pinyin, which can be laughable for a Japanese title that also includes kana. So please, please reconsider this problem and provide a solution that works for our multillingual world!