What's up タディちゃん
the Todd Rudick/Rikai.com blog
12/19/2004
New way of doing Rikai in the works
Inspired by google's neat javascript input completion code used in google suggest, I wrote a new version of the popup parsing code.
This could have huge advantages if it works out--first of all, I will be able to have popups for hiragana. The page sizes will be tiny, containing no additional markup. I don't need to segment Japanese, I let you do it by where you point the mouse. One disadvantage is the pages wouldn't work offline anymore--a quick call to rikai happens whenever you Mouseover a japanese character. Despite the massive number of callbacks this would create, I think overall it'd be less load for the server than parsing, exploding, and compressing the whole document the way I do.
Anyway, click the example above, and let me know what you think. It works in IE and Mozilla, I'm going to give up on Netscape 4, but I don't know if it works on other browsers.
This could have huge advantages if it works out--first of all, I will be able to have popups for hiragana. The page sizes will be tiny, containing no additional markup. I don't need to segment Japanese, I let you do it by where you point the mouse. One disadvantage is the pages wouldn't work offline anymore--a quick call to rikai happens whenever you Mouseover a japanese character. Despite the massive number of callbacks this would create, I think overall it'd be less load for the server than parsing, exploding, and compressing the whole document the way I do.
Anyway, click the example above, and let me know what you think. It works in IE and Mozilla, I'm going to give up on Netscape 4, but I don't know if it works on other browsers.
Comments:
I like it! Sometimes the katakana words can be hard to figure out, especially if they were not originally Enlish words.
Any plans to add some of the more, humm, interesting slang words to Rikai? Know of any good slang dictionaries J to E?
Any plans to add some of the more, humm, interesting slang words to Rikai? Know of any good slang dictionaries J to E?
# posted by 
: 5:45 PM
: 5:45 PM
Actually it always supported katakana to some extent, at least if it was just one word.
I don't mind having slang, and there is already quite a bit in edict already (which for now I don't filter out though it's marked). Anyway, it's easy to add a new edict-style dictionary to the set (which is currently edict, a name dictionary, and a place-name dictionary) if it becomes available.
I don't mind having slang, and there is already quite a bit in edict already (which for now I don't filter out though it's marked). Anyway, it's easy to add a new edict-style dictionary to the set (which is currently edict, a name dictionary, and a place-name dictionary) if it becomes available.
Todd. First, just a great big thanks for your work on one of the best Japanese sites on the Net. Love the site, the daily idiom, and all the rest.
Now, since you invited comments, I looked at the sample of your latest work and it does indeed read the hiragana too, which has huge potential. However, the definitions that came up for say は or といった were just wonky and bad. (は as a grammatical particle, topic marker, etc; といった as another functor indicating that something is called such-and-such.) All I got back were inapplicable kanji like 派 and 葉。
Also, the cute red box showing how the searched text was being parsed was absent! And (sorry to whine so much), it *did* seem to be slightly more time-consuming despite the initial time-save.
Nevertheless, Mr. Rudick (タディちゃん?) you have given us Japanese-users on the Web a wonderful resource and I add my two cents with much thanks and praise.
Happy holidays!
Anonymous in Hokkaido
Now, since you invited comments, I looked at the sample of your latest work and it does indeed read the hiragana too, which has huge potential. However, the definitions that came up for say は or といった were just wonky and bad. (は as a grammatical particle, topic marker, etc; といった as another functor indicating that something is called such-and-such.) All I got back were inapplicable kanji like 派 and 葉。
Also, the cute red box showing how the searched text was being parsed was absent! And (sorry to whine so much), it *did* seem to be slightly more time-consuming despite the initial time-save.
Nevertheless, Mr. Rudick (タディちゃん?) you have given us Japanese-users on the Web a wonderful resource and I add my two cents with much thanks and praise.
Happy holidays!
Anonymous in Hokkaido
# posted by : 3:41 PM
: 3:41 PM
Good feedback.
The handling of といった was a bug. I wasn't allowing conjugations of expressions ("exp" in edict), though of course they can end in verbs. There may be other cases, so let me know.
Not all small particals are listed in edict. I can add a few obvious one's (or so can anyone, see edict's docs), but of course in the end Rikai's for advanced readers and this isn't really going to help. Still, it looks bad when Rikai doesn't know such basic stuff. Anyway, the main advantage will be when nouns and verbs written in kana.
The red boxes are gone but highlighting is in. There's a purely technical reason for it.
The speed of the popups is definitely the negative. The big question is whether the extra little waits are worth being able to parse the document yourself starting at any character, plus getting kanji-cards in the popups? Is that useful/benificial?
From my end, it should be easier on the server overall, as I don't have to do any http compression or parse html--though I have to see how my hosting company likes all the extra tiny hits.
Post a Comment
The handling of といった was a bug. I wasn't allowing conjugations of expressions ("exp" in edict), though of course they can end in verbs. There may be other cases, so let me know.
Not all small particals are listed in edict. I can add a few obvious one's (or so can anyone, see edict's docs), but of course in the end Rikai's for advanced readers and this isn't really going to help. Still, it looks bad when Rikai doesn't know such basic stuff. Anyway, the main advantage will be when nouns and verbs written in kana.
The red boxes are gone but highlighting is in. There's a purely technical reason for it.
The speed of the popups is definitely the negative. The big question is whether the extra little waits are worth being able to parse the document yourself starting at any character, plus getting kanji-cards in the popups? Is that useful/benificial?
From my end, it should be easier on the server overall, as I don't have to do any http compression or parse html--though I have to see how my hosting company likes all the extra tiny hits.
