combine it into Android IME?

[c93...@gmail.com]

I like to use the english to kana tool to help look up information from Japanese sites. Is it possible to make a simple IME that when user type via keyboard, it can send string to the service's server and show corresponding kana in either candidate list or to the output? That would help users to look up things in a more simpler way.

[Ben Bullock]

On Thursday, 30 July 2015 07:38:08 UTC+9, Phake Nick wrote:

I like to use the english to kana tool to help look up information from Japanese sites. Is it possible to make a simple IME that when user type via keyboard, it can send string to the service's server and show corresponding kana in either candidate list or to the output? That would help users to look up things in a more simpler way.

I don't know how to make an Android application. When I let someone use kanji.sljfaq.org for the Android WWWJDIC, it was quite a big problem including causing the server to stop functioning due to badly-formatted requests the application was sending. Therefore I don't want to allow third-party access like that again. Sorry about that, but please use the normal web page. I'm working on redesigning it for mobile screens, but I only have a limited time to do this so I cannot give a deadline.

 

[tre...@gmail.com]

On Friday, July 31, 2015 at 12:24:00 PM UTC+10, Ben Bullock wrote:

I don't know how to make an Android application. When I let someone use kanji.sljfaq.org for the Android WWWJDIC, it was quite a big problem including causing the server to stop functioning due to badly-formatted requests the application was sending. Therefore I don't want to allow third-party access like that again. Sorry about that, but please use the normal web page. I'm working on redesigning it for mobile screens, but I only have a limited time to do this so I cannot give a deadline.

Can the code for the service be made public, or at least part of it? I have a largeish one-off integration I'm trying to do for a personal project where one of the last blockers is not having a way to read out English into katakana.

If the code can't, can enough details be released in order for someone to reimplement it similarly? I would probably be planning to rewrite it in whatever language I'd be using anyway... which might be Ruby, might be Java, not really sure because it depends on the other parts.

What I know from past posts:

- There is an English to IPA dictionary somewhere which uses British pronunciation. I'm sure I could probably dig this up if I tried...

- There is some kind of rule set for converting IPA to katakana. I think I could knock this out with regular expressions and lots of trial and error but it would suck.

- There is some kind of fallback logic for when dictionary entries are missing. This seems like the hard part. I am sure regular expressions would allow for it, but I wouldn't have the slightest confidence in being able to make it solid.

After mailing you directly from the address you published on GitHub, I got no reply. I since posted a more generic question about my problem on Stack Overflow:

http://stackoverflow.com/questions/33459414/can-mecab-be-configured-enhanced-to-give-me-the-reading-of-english-words-too?noredirect=1#comment54744176_33459414

The comment there suggested posting here, which is why I'm essentially repeat-posting, so sorry about that. :(

TX

[Ben Bullock]

No, sorry.

http://www.sljfaq.org/afaq/faq-faq.html#faq-faq-buy-e2k