Thank you mikankun!

Nukemarine Wrote:Here's the final file: Kore 2k/6k Optimal Indexes

Now, if someone can give me an unmunged set of audio and images, I can add those columns. After that, it's a simple matter to create the Anki files.

Has anyone made an Anki deck from the new optimized spreadsheet? That correspond to the media files and such, or are we still waiting on unmunged audio files (whatever that means ).

Yay an update. Welcome back Nukemarine.

rahsoul Wrote:Has anyone made an Anki deck from the new optimized spreadsheet? That correspond to the media files and such, or are we still waiting on unmunged audio files (whatever that means ).

I just uploaded a new Anki deck. The title is "Core 2k/6k Optimized Japanese Vocabulary" and is the entire deck. I'll edit the first post to refer to this deck as I'm sure having separate decks created confusion for beginning users of Anki.

This new deck looks awesome. Thanks Nukemarine! This is a classic thread.

Nice. I've only just started Core 400 so I've considering switching to this deck.

Currently my listening is a bit poor so I was studying kana + audio -> kanji + meaning. I was planning on doing this for the first few hundred cards at least.

Interestingly though your beginner cards are now English cloze deletions. Given how most on this site are opposed to studying English -> Japanese, what was the reasoning behind this?

Thanks for making such a great guide by the way.

*edit* one more thing - it seems to be missing card 6000? "An unexpected problem arose."

bchallenor Wrote:Nice. I've only just started Core 400 so I've considering switching to this deck.

Currently my listening is a bit poor so I was studying kana + audio -> kanji + meaning. I was planning on doing this for the first few hundred cards at least.

Interestingly though your beginner cards are now English cloze deletions. Given how most on this site are opposed to studying English -> Japanese, what was the reasoning behind this?

Thanks for making such a great guide by the way.

*edit* one more thing - it seems to be missing card 6000? "An unexpected problem arose."

Production reviews are the strongest reviews. If it wasn't for the cloze deletion sentences someone put the time to create, I'd have probably stuck with kana to kanji. However, kana to kanji is pretty much recognition where even the sentence and picture gave too much of a hint. Now you get English to Kana+Kanji which will pay off greater dividends. You have to know the words and hopefully there's enough clues to bypass legitimate but false results.

On the intermediate cards, there's merit to keep doing English to Kanji+Kana (fairly easy to switch cards to that) or to even create two sets that's both kana to kanji and kanji to kana cards.

Main reason I made the switch is realizing these decks, along with grammar and kanji decks, are meant to get you to the real Japanese material where it'll be all about sticking to as much Japanese as you can. The deck is simple enough that anyone that cared to can switch it up with minimal fuss.

I had been wondering for a while if I should go through the Core 2k6k deck. I've already done KO and would hate to waste my time going over vocab I already know, but I've found that my biggest issue is that even though I encounter words I "know" in the wild, I can never remember what they mean. I know that this is common and a natural progression of learning the words, but it's very frustrating.

I sampled a small batch of the production cards in this deck, and it might just be exactly what I'm looking for. I'll give it a whirl soon enough and see how it goes. Thanks for putting the deck together.

But all that yellow! I prefer to review with style.

So, I fancied it up just a little.
<image removed>

If anybody else fancies this layout, extract the images to the media directory. Then:
- Put this in the question template.
- Put this in the answer template.

See update: here.

Works on Ankidroid, too.

The background image is scaled proportionally and therefore cropped, and the deck goes up by 4MB since the templates are larger for each card, but you don't really care about any of that.

Netsplitter, how do you get your Anki interface to look like that? Is it through Windows or through Anki settings? And I really like the style

The colours? It's running on Linux, and that's a KDE colour scheme. KDE happens to use the same UI framework as anki (Qt4), so they play very nicely with each other. The button area colours wouldn't change though, and there's a plugin to fix that (descriptively called "Get answer button area background to use Qt color scheme")

I don't like staring at bright things, that's all.

Thanks, netsplitter!

Nukemarine: Thanks so much for the effort put into the decks and information on this thread! The new deck looks cool and just what I was looking !

Since I've been going through the cards with the photos showing (I used to hide them), I notice more than just a few are mixed up. If any one has a corrected list or something so that I can match up sentences with the correct photos, please post a link. I'll update the Anki shared deck soon after that.

Netsplitter, love the card layout. Only reason I choose yellow and black was it's similar to Koohii's layout which is relaxing to read over black letters on white background (for me at least).

I've gone through the first 500 cards of the new Core 2k/6k deck with English cloze deletions and I'm loving it so far. It's exactly what I need. I'll discuss a few changes I made to the layout from my previous post some time later this week.

For now, I'm wondering what you think is the best way for fixing errors in the deck. That is, where is the official master copy kept? Who edits it, and how? There are a bunch of errors in the cloze deletions (and other small things) which I'm fixing as I go along, but it'd be a waste to keep the fixes to myself.

I can give editing rights to the main list to a few that would like to update it. I haven't been correcting the main spreadsheet, just my own deck since I'm starting from scratch. Given that, no problem accepting help to update the google doc.

The cloze delete errors are annoying and a bit tedious to fix since it's three or four items on the fact. The image errors less so. For both, it may be best to tag the card each time an error comes up, then after a few are marked fix as a group if possible.

Anyway, send a request via e-mail from the google doc (I had two requests already I haven't got to yet).

Nukemarine Wrote:For both, it may be best to tag the card each time an error comes up, then after a few are marked fix as a group if possible.

That's what I've been doing so far for cloze/other errors, but I'm not bothering with the images. I'll leave that up to you. How do you know which image is the right one, anyway?

Nukemarine Wrote:Anyway, send a request via e-mail from the google doc (I had two requests already I haven't got to yet).

Is this an option in Google Docs? I'm not seeing it anywhere.

netsplitter,

The pictures are more of a "this picture has nothing to do with the sentence" then finding one that does seem close enough.

For google docs, send me an e-mail to my username at gmail.com.

A gentleman named CJon256 was able to automatically locate the mismatched images using an older list he had available. Definitely better that doing it manually for almost 300 images. Anyway, I've updated the main spreadsheet on google but I'm holding off updating the Anki file until the cloze deletes are finished.

For the clozed delete, it'll be difficult to use a spreadsheet formula to find the mistakes. Can anyone do quick method to compare the main vocab word with the sentence with the word highlighted with <b></b> then post those index numbers as the ones needing change. Likely just the first one or two characters should be compared to prevent false positives due to conjugation of verbs. That can give volunteers and myself the sentences they need to change without going sentence by sentence looking for them.

It should be simple to do with a python script. I'll whip one up now.

Edit:
I've decided that this will be the quickest way to do it:

I wrote a script that goes through my anki deck and checks for any discrepancies between the first character of the Vocabulary-Kanji field and the first character of the contents inside <b></b> of the Expression field. I have a list of 84 cards with errors now. Since I've already gone through the first ~900 cards and fixed the errors myself, about 25 of those are already fixed. Another 10 or so look like they have hiragana as the keyword but katakana in the sentence (do we want to fix those?), so they don't really count.

So my plan is to manually fix the remaining 50 or so this weekend. Then, use a bit of wizardry to fix the main spreadsheet. Of course, I'll test it on a private copy first, and I'll generate a list for you guys to look over just in case I did it all wrong. But I think this will be the quickest way.

netsplitter,

I really love your question and answer templates for the deck, but I just have problem. Kanji are replaced by their readings in the question container rather than displaying the furigana for it. I'm not sure what's causing it, but do you know how to fix that? Thank you.

I actually thought that was intentional, and that's how I've been using it. It makes you think a little harder, which is (strangely enough) just what I wanted, although I can see why others might not want it. I'll have a closer look at it tonight though since that field should be printing furigana in the first place. It worries me that I never noticed that.

I've also made a bunch changes to the template that I'll post later tonight. Sorry for the delay in submitting the cloze fixes as well. I ran into a few issues (and insist it's the computer's fault). I'll have those done in a day at most.

Dashwolf, it turns out that that's how the Japanese plugin does things. It removes kanji from question cards and replaces them with their readings. Not much I can do about it (although I disagree that it should outright make this decision on your behalf). Ankidroid mimics this behaviour as well.

Now, about the template. It looks like this now. I've removed some of the excessive padding and the "notes" field, which was useless (to me). Since I review on my phone, I need as much space as I can get.

I also centered the image properly and fixed an error where it breaks on apostrophes in the English meaning.

I changed the English cloze portion to use a monospace font, and the Japanese text is using my font of choice, VL PGothic. I've removed the font change from the template I'm posting here though since I know people are very particular about their fonts. You can still change it from the field editor if you want. (The default Japanese font on Android uses Chinese variants of a lot of the characters, which is very frustrating, so if you use Ankidroid I encourage you to change it.)

The only major change is the English sentence, which I am hiding on the question side. I found the English sentence very distracting; it's the first thing my brain notices and it won't let me read the Japanese text without sliding over to the English and reading it until the end. It also meant putting much less effort into trying to derive a meaning for the sentence since I've already read a translation of it. I can still show the English by hovering (or tapping) the covered portion if I need to. Sometimes this is required since it isn't obvious what the sentence means with a missing word (usually at the end). The English sentence is revealed on the answer side.

I also added a hint field, which is sometimes needed to disambiguate similar terms. (I'm adding these for myself and aren't part of the original deck).

As before, if you want this layout, make sure you have the images in the media directory and:
- Put this in the question template
- Put this in the answer template

Of course, it's just a template, so feel free to modify it to suit your needs where they differ from mine.

netsplitter,

Thanks for your help and updates. Hiding the English sentence with option to hover over it is extremely helpful, because I was having the same problem with it. I've also been trying to find a nice font for Japanese text and VL PGothic looks great.

I'm done with fixing the cloze deletion errors. I think comparing the changes is a little unwieldy, so you might just have to trust me that they're correct. There's a diff over here if you want to have a look, or download a copy of the output and compare them with your own diff tools. Or just take my word for it.

Nukemarine, I will submit the updates to the main spreadsheet if you're fine with those changes.

There are also lots of harmless formatting issues that I'd like to fix eventually. Things like extra spaces or commas inside furigana (e.g., 明日、 [あした] instead of 明日[あした]、 ). It misaligns the furigana, and that upsets me . It's mostly automated fixes so it shouldn't be a problem.

Edit: I updated my own deck with the new Sentence-Image paths, and now they're all wrong. Either the list you used was bad or you added it to the spreadsheet incorrectly. Please look into that.

Double edit: It looks like the corrected list was indexed against the opt-voc-index instead of the core-index field. That should be an easy fix. I'll do it now.

Sorry for the stupid questions but I'm kind of confused on how to begin with the first 555 in RTK Lite.

From what I understand the 555 is based on the first 555 in KO2001, so I should ignore Heisig's ordering and jump around in the book according to KO2k1? Or follow Heisig's ordering but skip non-KO2k1 kanji? Or should I not be using the RTK book itself at all, and just using Anki with Heisig's mnemonic system/stories from here?

Dashwolf Wrote:netsplitter,

I really love your question and answer templates for the deck, but I just have problem. Kanji are replaced by their readings in the question container rather than displaying the furigana for it. I'm not sure what's causing it, but do you know how to fix that? Thank you.

Here's how you can fix this

1. Settings>Plugins>Open Plugins Folder
2. Open the 'japanese' directory
3. Open 'furigana.py' in a text editor.
4. Search for 'reading_in_question' and change the 1 to a 2. You can also set it to 3 to show the kanji and have the reading in a tooltip.