If it's using Google Transliterate, that's deprecated
https://developers.google.com/transliterate/
The question is whether or not there is a suitable replacement for it. translate.google.com has this feature, not sure if there is a (free) API for anyone to use though.
And you should be aware that bringing a third party into the editor, may raise privacy concerns. Anything you type has to go to the Google server and back. At least that is how it works with Japanese IME as it must be backed by a huge dictionary to look up words in so if you type kanji, it can offer you 漢字, 監事, 莞爾, and other words which may fit that reading.
If it's just a matter of simple replacements, like a purely phonetic script w/o dictionary, such as replacing a with あ, na with な, ta with た and so on and so forth, you could do this with a bit of JavaScript yourself, not requiring a third party at all. I actually did this for Japanese kana in the MyBB editor.
oh yes, a phonetic script would be great but may be it will not be accurate. As far as my experience goes, google transliterate was very accurate. Thanks for the reply
well, which language or script do you need, specifically?
i need for telugu. This is my website.
www.teluguwebworld.com. I want to develop a language pack too, since there is no telugu language pack for MyBB
. Can u help me in this regard?
learnt that google has a pricing for its api
. 20$ per million chars..
Not sure if I can help. I could provide you with the code for Japanese.
Basically it contains a mapping like so (excerpt):
...
"e": "え",
"n": "ん",
"pi": "ぴ",
"tsu": "つ",
...
And in the editor you write ascii and select it and then press the kana button. Or you press the kana button without selecting anything and type into an input box. And if you type enpitsu it turns it into えんぴつ. There are also longer and shorter syllables, it looks for the longest one in the mapping first.
Now the question is whether the same method can be applied for telugu or not, and how to come up with such a mapping that covers all the characters, and how large such a mapping would be since the users have to download it as part of the JavaScript.
For Linux IME there's a project called m17n which seems to contain a file te-rts.mim which seems to contain the transliteration rules, but I'm not entirely sure as to by which rules this data is applied and how to duplicate that in JavaScript.
If you're lucky you could google if anyone has solved this already independently, like a JavaScript that lets people type telugu in any textarea. Then you could use that in the editor without actually modifying the editor...
