24 Aug 2011 00:17
I'd like to take a crack at adding word-frequency support to the source. Any hints?
George Michel <sarumonakuba@...>
2011-08-23 22:17:27 GMT
2011-08-23 22:17:27 GMT
Despite it's age this program still seems to be the best at what it does. I think only Wakan offers similar functionality, though not nearly as well. Many thanks Mr. Koch.
I've been using it with Bullwhip PDF printer to output fully annotated Aozora PDF's to my Kindle, which has been an INCREDIBLE boon to my reading of difficult texts. For a better reading experience I've had to meddled around with Edict a little, eliminating superfluous entries and parentheses, as well as facilitating tategaki-style pages with a rotated version of MS Mincho (as well as dummy versions of Mincho and Arial for annotation-free pages interspersed between the annotated ones)
It'd be ideal to implement a lot of this additional functionality in code -- ie. allowing access to the <at> fonts to obviate the need for custom tategaki fonts, exporting to PDF with proper page breaks via itextsharp or a similar library, an option to hide the annotations without changing the line spacing -- but I doubt I'd be the man to do it, as I'm still very much a beginner in Java.
The one thing I WOULD like to take a crack at changing is the way JGloss chooses its translations. I'm using ChaSen and EDICT.. Two syllable words not written in kanji in the text are invariably wrong (ie. "tachi," amusingly, ends up as "the dominant partner in a homosexual relationship," rather than "group" for instance). The jjdx index orders everything in alphabetical order, and seems to simply choose the first reading, katakana first.
Incidentally, I've managed to find a frequency-ordered version of Edict packaged with a program called cbjisho in sqlite format, which I then converted to a plain text Edict-formatted dictionary, if that couldt help. Unfortunately the index-building classes seem to quicksort any indexes based on it back into submission. Telling FileBasedDictionary.class not to quicksort, predictably, simply messed things up.
------------------------------------------------------------------------------ Special Offer -- Download ArcSight Logger for FREE! Finally, a world-class log management solution at an even better price-free! And you'll get a free "Love Thy Logs" t-shirt when you download Logger. Secure your free ArcSight Logger TODAY! http://p.sf.net/sfu/arcsisghtdev2dev
_______________________________________________ JGloss-discussion mailing list JGloss-discussion@... https://lists.sourceforge.net/lists/listinfo/jgloss-discussion