Thanks guys, that helped a lot. It seems tags are what I was looking for, however the models are much more flexible than I previously thought. So now I have the new problem of figuring out how to use this power to configure my deck. Thanks

Anki is a tool that has to be learned in its own right. To anyone starting an anki deck, I recommend not to start too fast. Experiment a lot until you are sure, what models, decks, cards, facts, tags, card models etc. are and how they work. You might run in danger to end up with a totally messed up deck. I personally had to restart my deck two times until I ended up with "clean" decks (no extra models from importing etc) and card models that do exactly what I want them to do.

misha Wrote:Importing RevTK progress into Anki 0.9.x plugin is now available at: http://ichi2.net/anki/wiki/Plugins

Let me know if there are any problems.

When I imported all of my progress, both my failed stack and all of my due cards became "failed" in Anki. Is this normal? As new cards become due, they also are showing up as failed instead of awaiting review.

I've done a few hundred reviews since importing them into Anki, so maybe there's something in the database I can change at this point to fix it? It's annoying since it looks like all of the ones I'm actually failing while doing reviews won't come back until I clear out several hundred cards that are really just overdue.

How do you view the plugins folder? Idiot's guide please.

Hey, I was having the impression anki would find the readings of the kanjis for me. But all it does is give me the options. Is it right or it is not properly configured ?

I think I've solved the problem, but I'd definitely appreciate feedback from someone who knows the database structure better. (Is there any documentation for it?)

It looks like the cards are showing up as "failed" because the plugin never sets the successive field. I ran the following queries:

update cards set successive=7 where successive=0 and firstAnswered=0 and interval=240
...
update cards set successive=1 where successive=0 and firstAnswered=0 and interval=3

(Seven in total, substituting in the proper interval for each box number.)

The deck now *seems* to be working as expected.

kaiya Wrote:How do you view the plugins folder? Idiot's guide please.

What OS are you running?

Vista: C:\Users\<username>\.anki\plugins
XP: C:\Document and Settings\<username>\.anki\plugins

Scout,

I notice your algorythm sets definate days per box (3, 6, 14, 30, etc.). Although it's not a big problem, the RevTK gives last review time and next review time. If you use that to set review length, it mimics better what RevTK has set. Plus, if you look at the cards due by graph, it'll be a lot smoother instead of 8 spikes.

Anyway, I've had problems using the new plug-in, and I've procrastinated figuring out why. Basically, I created 2 cards (kanji to keyword, keyword to kanji) and wanted both to have the same due times. It did not recognize the attempt.

Those numbers were actually set by misha's plugin, based on the Leiter Box number that was in the exported data. (I'm just changing the successive field.) misha's script does use the current RevTK expiration date. The interval information seems like extra data in the Anki database -- for stats maybe?

It looks to me like the plugin should be changed to set the 'successive' field to the Leitner Box number that RevTK includes in the exported data. (My queries were just a hack to get that data set correctly without re-running the plugin and overwriting the reviews I'd done in the past few days.)

For those interested, I also imported my stories from RevTK. It's definitely a hack, but I'll probably post the details after I clean it up, and make sure that it didn't cause any problems.

mentat_kgs Wrote:Hey, I was having the impression anki would find the readings of the kanjis for me. But all it does is give me the options. Is it right or it is not properly configured ?

It's right. You need to select the proper reading when there are 2 or more possible readings for a word. In 0.9.6, you can select the proper reading just by clicking on it. (Which is totally awesome, resolve. Best new feature EVER!)

rich_f Wrote:In 0.9.6, you can select the proper reading just by clicking on it. (Which is totally awesome, resolve. Best new feature EVER!)

My thanks as well!

I just wish there was some way to lock the autogeneration feature. I actually stopped putting the reading field on my cards, and just copy and paste the contents of it into the top of the meaning field. I got really sick of fixing typos in the expresion field, only to realize that I now had to go back and re-edit the reading field for the right reading choices.

Well, I have to admit that I look forward to adding cards now that you can click on the readings you want. Makes it bucketloads easier.

This may be a silly question, but does anyone know where I can find the kanji report statistics on anki? I've looked everywhere but I have no idea. Thanks!

It's under the tools menu.

I can't believe I missed that... thanks very much!

scout Wrote:

kaiya Wrote:How do you view the plugins folder? Idiot's guide please.

What OS are you running?

Vista: C:\Users\<username>\.anki\plugins
XP: C:\Document and Settings\<username>\.anki\plugins

Thanks.

If it`s not on the to do list, there should be a way to suspend auto-generation of reading. I removed mine as it just was not worth the time loss.

I'm sure you could just edit the card model. Remove what's in the "Special features used by this field"

I think he's saying that's what he did, and that something like a button toggle would be more convenient.

You know you can double click the readings to choose the right one, right? After you've added text once, why would you want to change it? I can't say I've ever felt the need to edit my own cards after adding them more than a few times.

I usually have to change it because I find typos. This doesn't happen *that* often, but often enough that I always copy the final reading to the "meaning" field and don't have the reading field display on the cards.

Being able to double-click on a reading in 0.9.6 is a huge improvement, but kakasi still makes mistakes at times. That means that every time it gets re-generated you have to choose the right readings, and then read through the whole thing very carefully to make sure everything else is correct. This is an bigger issue if you make a change a few weeks after you enter the card (e.g. you want to add a line break, or a period, or fix a typo) and it's no longer easy to find the source for the original sentence to double-check that you got all of the reading choices correct. Granted, this really wouldn't be a real issue for a vocabulary word deck, but it's significant in a sentence-based deck.

This sounds like it wouldn't be a problem if you double-checked your sentences before saving them?

I'd estimate at least 3/4 of these changes happen while editing the card the first time. Usually it's enter sentence -> fix reading -> add stuff to meaning side -> oops, realize I made a mistake in the sentence. It's usually one kana, so for me it's much easier to just fix the single mistake in two places than to have to bother re-selecting the correct readings.

edit: I think for me, I'd like to completely turn off auto-generation of the reading field, and instead have it done with a key-combination similar to how Ctl-enter adds a card.

Why don't you change your workflow so that you proofread the expression before editing the meaning? :-)

The easy answer is I do proofread it at the same time, but I occasionally miss things. There are also other reasons to change the reading side. These problems essentially stem from entering stuff directly into Anki before finalizing the card. (I've tried putting stuff in a text file first, and making cards later, but that basically means the cards never get made.) There are times where only after I've gone through all of the definitions for words in the sentence do I realize something is ambiguous and I want to get rid of that part of the sentence (which regenerates the meaning.)

Stepping back, the 'right' thing to do would be select the readings after everything else is done, but that's a bit tedious if you are taking your sentence from printed material, where you have to hold the book/magazine up while typing in the sentence. It's much easier to get the reading entered at the same time.

PS: See my edit above for my suggestion of how I'd like the reading field generation to work.