darg_sama and I have put together an Excel vocabulary drilling tool (actually an extensive modification to ocha_no_kanji under the GPL) at http://www.japaneselearningtools.com that we would like people to beta test.
.

Some of the nice features are
* Covers all Kanji vocabulary words from all JLPT levels. We will be adding all JLPT non-Kanji vocabulary shortly.

* Uses a modified Leitner fading (uses a Leitner-weighted random selection function to reduce predictability and allow you to review as many words as you want any time you want).

* Simple to use. Just press "Start Test" for randomized flashcards.

* Checks pronunciation entered in hiragana or virtually any Romanization system including Hepburn and Kunrei-shiki.

* Appropriate for all students as it progresses in JLPT and grade order.

* Automatically displays your progress over time in Excel charts

* Free software under the GPL.

If you're interested and have Excel (and adventurous enough for beta testing), go to http://www.japaneselearningtools.com and give us some feedback.

First of all, I really like this idea. I hadn't found any vocab tool I liked so far, and this seems a good one. I've tried a few rounds and I have noticed the following:

- 一月 without meaning pointer I typed 'ichigatsu' which it told me was incorrect... This vocab does not distinguish clearly between ichigatsu and hitotsuki with a meaning pointer.

- Bug/runtime error: after I'd done a round of 50, I tried to do another one. The colour came up grey instead of red. That looked odd, so I pressed the x to get out of that screen. Afterwards, I wanted to start up again, but it gave a runtime error '380', could not set ListIndex property. Invalid Property Value.

Anyway, all was well when I saved the spreadsheet, went out of Excel and restarted the whole thing.

Anyway, thank you, great idea, keep up the good work!

This is G-R-E-A-T!! Thank you very much for putting this together. I've spent about 45 min. on it today for the first time and found no bugs or problems.

Well I did find something else a little awkward: when entering 'otoko no ko' it did not recognize it as correct, and I had to enter it as 'otokonoko' instead. Once I knew, I could avoid it, but it would be nice if both were recognized as correct if possible.

Just a small comment, and I hesitate to say anything because I know you have put so much time and energy into this project, but the red screen is a bit difficult for me to use. I think it's because of the black lettering...perhaps the lettering could be changed to white? or the background color changed to something a bit easier to see? I try to hurry through the red screen to the next screen (I think it's a light gray) because it's like going from night to day. I'm a bit of an "older" student so one factor could be that my eyes are a bit more sensitive than young people's. Anyway, just a comment. Perhaps the user could control some of the colors?

This is an excellent learning tool. I am going to try to get it to work on my Palm. I have been using twinkle on my palm for all of my current vocabulary study, but this is a very good tool as well. I am going to pass it on to some other friends who are learning Japanese.

Thanks for all of the hard work guys.

This is definitely an awesome tool, it seems like it will be another great addition to my Japanese learning resources. Many thanks for putting the time and effort into creating it!

I seem to have found a small (on my system anyway), when a suffix question appears (～人、～日), it seems that no matter what I type, ～にん、ーにん、にん、whatever it tells me I am wrong.

Has anyone else found this problem?

Christoph Wrote:I seem to have found a small (on my system anyway), when a suffix question appears (～人、～日), it seems that no matter what I type, ～にん、ーにん、にん、whatever it tells me I am wrong.

Has anyone else found this problem?

Me too. I got an email from mspertus, he reckons this could be to do with entering the answers in kana, which I was too.

Agree that is looks good and will hopefully get even better.

Cheers,

do not include the ~ in your answers

As for entering "otoko no ko" vs. "otokonoko" the Japanese don't write with spaces between words so that is why the program only accepts the "otokonoko" rendering. At least that's my two-cents.

Johnnyltn Wrote:do not include the ~ in your answers

It replied "incorrect". This is when entering the answer in kana.

Cheers,

Oh, sorry, I didn't realize about the kana thing. I'm entering using romaji and everything seems to be working fine.

OK. I've put up a new version that solves the "~" problem and allows spaces in romaji. Let me know if that resolves the issues. (Unfortunately, you will lose your history. Our next major update will have the ability to import your history, so this "should" be the last time you lose your history.)

Mike

astridtops Wrote:First of all, I really like this idea. I hadn't found any vocab tool I liked so far, and this seems a good one. I've tried a few rounds and I have noticed the following:

- 一月 without meaning pointer I typed 'ichigatsu' which it told me was incorrect... This vocab does not distinguish clearly between ichigatsu and hitotsuki with a meaning pointer.

I agree it looks a very promising tool. Thanks for all the effort!

I also got confused with 一月 without a meaning pointer.

Johnnyltn Wrote:Oh, sorry, I didn't realize about the kana thing. I'm entering using romaji and everything seems to be working fine.

Yup, when I started entering answers in romaji, everything went smoothly. There was no need to enter the '~' symbol. However entering kana is still problematic in this regard.

I think it would be good if the text input box held the answer that you type rather than changing to the correct answer. As the answer is displayed below it is redundant AND it would be nice to compare what I type with the correct answer...only on the very rare occassion I get them wrong


Regarding how the testing works, if you pass say 45/50, you then go straight on to the 5 that you got incorrect and it keeps testing you till you get them right. I assume that is how you designed it, but I just wonder if that is the best design. I have no real Leitner knowledge other than a basic understanding of the principles...BUT for me personally, I just like to move on.

So, say I had 5 failed kanji in a test I would like to be able to move on to the next 50 card test which could potentially contain the 5 failed from the previous test. When you go straight to your failed kanji, as you've literally just failed them, the answers are still fresh in the mind. So my feeling is that this doesn't really help push them into your long term memory.


Is there a way to skip the earlier tests and get to the kanji you are most interested in? I'm going for JLPT2 this year and as time is short I'd just like to skip to the kanji most relevant to me (at this stage in time).


'Excel@Japanese Help > Help' - it gives an error message.


Quit out of test then try to restart a test gives an error 380 > (details > Excel would crash (although this could just be my machine)) - restarting the sheet sorts it out.


Cheers,

mspertus Wrote:OK. I've put up a new version that solves the "~" problem and allows spaces in romaji. Let me know if that resolves the issues. (Unfortunately, you will lose your history. Our next major update will have the ability to import your history, so this "should" be the last time you lose your history.)

Mike

Nice one Mike.

That sorts out the issue with the '~' character. Typing the answer in katakana with OR without '~' works.


Cheers,

synewave Wrote:

astridtops Wrote:First of all, I really like this idea. I hadn't found any vocab tool I liked so far, and this seems a good one. I've tried a few rounds and I have noticed the following:

- 一月 without meaning pointer I typed 'ichigatsu' which it told me was incorrect... This vocab does not distinguish clearly between ichigatsu and hitotsuki with a meaning pointer.

I agree it looks a very promising tool. Thanks for all the effort!

I also got confused with 一月 without a meaning pointer.

Johnnyltn Wrote:Oh, sorry, I didn't realize about the kana thing. I'm entering using romaji and everything seems to be working fine.

Yup, when I started entering answers in romaji, everything went smoothly. There was no need to enter the '~' symbol. However entering kana is still problematic in this regard.

This is now fixed. See my last post.

synewave Wrote:I think it would be good if the text input box held the answer that you type rather than changing to the correct answer. As the answer is displayed below it is redundant AND it would be nice to compare what I type with the correct answer...only on the very rare occassion I get them wrong


Regarding how the testing works, if you pass say 45/50, you then go straight on to the 5 that you got incorrect and it keeps testing you till you get them right. I assume that is how you designed it, but I just wonder if that is the best design. I have no real Leitner knowledge other than a basic understanding of the principles...BUT for me personally, I just like to move on.

So, say I had 5 failed kanji in a test I would like to be able to move on to the next 50 card test which could potentially contain the 5 failed from the previous test. When you go straight to your failed kanji, as you've literally just failed them, the answers are still fresh in the mind. So my feeling is that this doesn't really help push them into your long term memory.

I've used it both ways with the failed kanji and prefer when it repeats until you get it right. However, this isn't imposed on you: The retesting is not stored in history, so this allows you to do it either way (at least once I fix the "Quit Test" bug you mention below).

synewave Wrote:Is there a way to skip the earlier tests and get to the kanji you are most interested in? I'm going for JLPT2 this year and as time is short I'd just like to skip to the kanji most relevant to me (at this stage in time).

No. For rather complicated reasons that darg_sama and I discussed in depth, we switched the program from being kanji-based to being word-based (We also each spent hundreds of hours using the kanji-based version). I plan to eventually write up the reasoning and put it in an onsite essay. Targeting specific kanji would be complex (but not impossible), so I'm not in a hurry to implement it. Fortunately, because of the Leitner algorithm, it will recognize that you know the basic vocabulary and only rarely test you on it, so it shouldn't take long to get to words you find a challenge. If you'd prefer to use the kanji-based version and manually select which Kanji to test, PM me for it and I can send it to you (it may not save you time, though, because manually selecting the kanji quickly becomes cumbersome).

synewave Wrote:'Excel@Japanese Help > Help' - it gives an error message.


Quit out of test then try to restart a test gives an error 380 > (details > Excel would crash (although this could just be my machine)) - restarting the sheet sorts it out.

Thanks for the great feedback!

Mike

Mike,

Question regarding Japanese@kanji...Word based as opposed to kanji based testing, I get that. Does it work from JLPT4 words through JLPT3 then JLPT2 towards JLPT1? I assume so but just wanted to clarify.

Cheers,

Seems to be doing just that. I spend a good part of the day yesterday testing more and more sets of 50, and suddenly, my nice green kanjis of the first two rows turned white and yellow. I was a bit alarmed at first, but then I realized I had crossed from JLPT4 to JLPT3 and was required to do more meanings for the kanji I already learned for JLPT4 meanings. Since my own profiency is somewhere around JLPT3 (we're not in the test loop around here), you'll have to figure out the rest yourself.

Oh, I did notice something else:
My kanji started to fade to light green after a mere 12 hours. Now, that could be because it doesn't take absolute time in account, but rather the amount of practice you put in since you last visited that particular vocab, but it annoyed me a bit that following 10 tests or so didn't include the fading kanji vocab. Also, I would hope that if I start to slow down and do only 1 or 2 tests a day, it would take more time for the kanji to fade?

Also, one of the things that I wonder about is that the program decides to progress at its own pace. Probably because I still get about 70% right most of the time, but sometimes I would like to tell it "no new stuff at the moment, give me a couple of rounds of recap instead". Maybe a system where you can choose between automatic progress (as it seems to be doing now) and consciously adding more vocab would be nice for those people who like to drill their existing vocab very thoroughly before moving on to new stuff.

I found another ambiguous one: for 十分 it didn't accept juppun, but juubun instead, so I think that needs a meaning pointer as well.

Just adding that I'm still very happy with the new tool, and I cannot applaud your efforts enough (despite all my critical remarks, which I do hope come across as having a constructive tone).

Do I have to start with all those level 4 stuff first? I'm aiming for level 2 this december. It's kinda a waste of time if I have to go through all level 4 and 3 vocabulary first. Great tool by the way, seems really usefull for all levels.

mspertus Wrote:

synewave Wrote:Is there a way to skip the earlier tests and get to the kanji you are most interested in? I'm going for JLPT2 this year and as time is short I'd just like to skip to the kanji most relevant to me (at this stage in time).

No. For rather complicated reasons that darg_sama and I discussed in depth, we switched the program from being kanji-based to being word-based (We also each spent hundreds of hours using the kanji-based version). I plan to eventually write up the reasoning and put it in an onsite essay. Targeting specific kanji would be complex (but not impossible), so I'm not in a hurry to implement it. Fortunately, because of the Leitner algorithm, it will recognize that you know the basic vocabulary and only rarely test you on it, so it shouldn't take long to get to words you find a challenge. If you'd prefer to use the kanji-based version and manually select which Kanji to test, PM me for it and I can send it to you (it may not save you time, though, because manually selecting the kanji quickly becomes cumbersome).

Same as you Renske, but let's face it, it isn't going to do any harm to go through stuff we should know.

I have been meaning to get started, but what I'm planning on doing is getting the 'List of kanji appearing in past tests' from the MLC JLPT2 guide http://www.mlcjapanese.co.jp/Download.htm into Mnemosyne. If you fancy giving me a hand inputting them, that could be our contribution. What do you think?

For the record, it took me about 2 hours this morning to get through the JLPT4 list with only an occasional mistake. Taking a deep breath and plowing through seems to be the quickest. Of course, it should be possible to swap spreadsheets and work in turns or something like that, if the mastery of the first levels is close to 100% it shouldn't matter who puts what in.

Thanks for all the great feedback!

The system should be self-adjusting. Give it a week, Synewave and Renske, and if it isn't primarily testing you on level 2 words, let me know and I'll add a "minimum JLPT level feature".

Astridtops, If you want to know why a kanji has faded, just click on it and type Ctrl-i (or choose "Show Kanji Info" from the toolbar). Once you've gotten a kanji right twice, it will fade a lot more slowly (and be tested less frequently). I'd also ask you to give it a week, and if it's still adding words at a high rate, let me know and I'll add a "No new words" feature. If you're scoring about 70%, it will stop adding words soon.

We did at least discuss the ability to select which JLPT level you're on and start by testing all words up until your selected level - would that be hard to add as an option? I can at least imagine how the logic would work, so I know it's possible. That is how it was originally, back when it was testing you based on kanji not words.

PS: starting on the non-kanji word list right now... it's pretty long, so expect a bunch of additional words to be tested.

I know you guys have your hands full but do you have any thoughts concerning the colors for the background? I had mentioned about the difficulty seeing the lettering on the red background testing area. Just thought I'd ask again in case this question might have slipped by.

Thanks again for the wonderful program.

If you release a new version, can I keep my current results somehow?