I wanting to start working on the Smart.fm Core 2000 list, with the wonderful Anki as my SRS, and I've run into a dilemma. What order should I learn the facts in? I'm using the Core2000 shared decks, by the way.

Should I just do use "Show new cards in order added" in Anki, or should I work on each word first and then work on the sentences with those words after that? The smart.fm lists seem to be ordered in the latter method, with you learning each word and then working on sentences later. Also, when doing sentences, should I mark it wrong if I don't understand the entire thing, or only if I don't understand the highlighted word?

Any advice I can get on this would be greatly appreciated.

Well, ultimately you'll have to experiment. I can share what I'm doing, but that's no guarantee that it'll be best for you.

I'm learning sentences only; no vocabulary-word cards. Also, my leech threshhold is set very low (6 failures) so I don't spend time on difficult cards to the detriment of easy ones. That trims about 2-5% of the most-difficult, so I could actually set the threshold even lower.

My review options are set up to be "greedy:" show the card I most need to review now. "Review from largest interval," and "New cards after all other cards." Failed cards show up after three minutes. Ideally, I'd like that to be "at least three minutes and after all reviews," but I don't know how to make Anki do that.

New cards graded Hard or Easy show up within 24 hours--a small initial interval, but I decided I prefer that over trying to set-up a mnemonic system. (I was spending too much time obsessing over mnemonics; not enough just doing it.)

Finally, I do "audio -> comprehension" first, then "reading -> spoken" cards. Minimum spacing is set to 5 days (7200 minutes) to ensure I'm familiar with hearing something before I try to say it.

I have a baseline goal of 30 minutes per day, but if I want, I'll keep going. I don't worry about how many new cards I learn or even whether I finish reviews--those things take care of themselves.

So, if you want to copy someone else's setup to start, there it is. Again, it's most important to just do it. Also, I think (personal opinion) it's pretty much required to do immersion along with Core2000: Japanese iPod, Japanese TV, Japanese News, etc. You don't necessarily have to take stuff apart with a dictionary (do it if and only if you feel like it), but you must show up and get the exposure. Otherwise, study methods like Core2000 are like planting a garden in sand.

brandon7s Wrote:I wanting to start working on the Smart.fm Core 2000 list, with the wonderful Anki as my SRS, and I've run into a dilemma. What order should I learn the facts in?

the first 2 steps are the easiest and step 3 is the least interesting, but other than that there doesn't seem to be any order

if I was doing it over I'd sort the sentences by kanji to limit the rate of introduction of new readings

cangy Wrote:

brandon7s Wrote:I wanting to start working on the Smart.fm Core 2000 list, with the wonderful Anki as my SRS, and I've run into a dilemma. What order should I learn the facts in?

the first 2 steps are the easiest and step 3 is the least interesting, but other than that there doesn't seem to be any order

if I was doing it over I'd sort the sentences by kanji to limit the rate of introduction of new readings

Someone wrote a little program that appeared able to sort words based on all the kanji in word, or at least tried to. I had hoped that program could then be used to sort words based on the KO2001 order.

Idea being, sort the Steps 3-8 Core 2k vocabulary list 401 - 2000 into an easier order to digest (Steps 1 and 2 are easy in any order). That would put all those "more common" (in a newspaper at least) but difficult words near the end instead of the beginning.

However, the sorting program didn't really come about, so the only order we have is the order based off newspaper frequency. Nearest substitute at the moment is Hashiriya's KO2001 group list at Smart.fm. I say Hashiriya's list cause he inputted more than half of the entries into Smart.fm.

Oddly enough, I'd say start with Step 10 and work backwards. The words are put in kana order but still much easier to learn in bulk than if you started with Step 3.

Have you thought about the 2001 Kanji Odyssey compiled lists on smart.fm? I started doing Core 2000, then switched to the KO, because of the order the of kanji compounds. I think they use the same smart.fm sentences as Core 2000. My method is to do sentences. I import a list and suspend them. Then I unsuspend 5 or 10 a day, in the morning, after I have cleared my reviews. I then go over the new cards, making MP3's for my ipod, and learning the meanings of all the words in the sentence. If I have time I grab better dictionary definitions, and edit in example sentences. Sometimes I can pass the new sentence in this first session, but usually just go over them to learn them, and leave them in Anki for later. Then I go about my day, with my ipod to go over the new sentences throughout the day. Then in the evening I clear the reviews again. I only pass if I understand the sentence, and all the words it contains, (and can pronounce it). Sometimes this means learning 4 new words in a sentence, but usually it is only one, sometimes two. I'm still finding my rhythm with it. I shadow during the reviews and during my ipod listening sessions, and try and feel the meanings of the words. Not sure if this is "the best" way to do things, but it seems to be working for me.

I keep a separate deck for the auto import smart.fm KO2001, and another for all else.

Hah - Nuke chimed in while I was writing my post (he deserves his share of credit for the KO list too).

Just want to add, I did Core 2000 for a while, but I find the KO order much better, as I get an ear for the readings with the compounds blocked as such, and it makes the "production" part (pronouncing the words - for me the hardest thing to remember) come easier I think.

Thanks everyone for your input, y'all have definitely given me something to work with here. It seems like everyone has a different way of going about these lists, so it looks like I'll have to experiment and find out what works best for myself.

I AM still working on memorizing kanji, but since I'm getting so tired of doing nothing but kanji, I figured I'd start on something else to keep it interesting. Also, since I'm not using the actual RTK list, I can simply add whatever kanji I run into while doing Core2000 to my kanji pile. Actually seeing the kanji I'm memorizing being used will help me memorize them better, I believe.

I do have a KO2001 deck, but honestly, I like the audio in Core2000 much better, which is mostly why I decided to start on Core2000. The order of the kanji compounds that they teach me doesn't really make a difference to me - so far I have no found that some are inherently easier than others to learn, but maybe that's due to a lack of experience.

Nukemarine, why do you think that it would be easier to start at Step 10 and work backwords through Core2000? What are the advantages to this?

IceCream Wrote:There's a Core series in the same order as KO2001. So, it's the Core series sentences and audio, just with a different order. That's what people meant rather than doing KO itself...

Oh ok, I thought folks were talking about actually doing the KO2001 list, not just the order. Does anyone have a link to this Core2000 in the KO order? I'd like to check it out. I did a brief search for it on smart.fm but didn't find anything - given more time I could find it, but if anyone has a link that'd make it much easier for me.

IceCream Wrote:its not that some kanji are inherantly easier to learn than others... its more that learning a bunch of compounds that have the same kanji in at the same time will help you recognise the kanji in other compounds, have a pretty good idea of what the reading will be. If you do them randomly, it makes getting to know all the connections between them a bit harder, i think...

I see now, and I think you're right - I'll have to give that a shot.

http://smart.fm/lists/browse?keyword=kanji+Odyssey

That will get you going, and yes, they use the same audio sentences that you like.

And I can't understate how big a help it is, in that the readings show up in patterns that reenforce each other.... imho much more efficient.

TaylorSan Wrote:That will get you going, and yes, they use the same audio sentences that you like.

And I can't understate how big a help it is, in that the readings show up in patterns that reenforce each other.... imho much more efficient.

Thanks for the link, Taylor - I'm going to import the 2001.Kanji.Odyssey Compiled List 01-02 and see how I like it compared to the standard Core2000 list.

Sure thing. I didn't go far in Core2000, but I noticed the sentences in the KO lists used many of the sentences I had done in Core. I'm guessing you'll get all the stuff from Core, but with the added bonus of order. I could be wrong about that though. Either way there is a lot of material there. It's kind of boring at first (I hate numbers) but keep at it, it does get better. Let me know how it works for you.

頑張って！

Nukemarine Wrote:

cangy Wrote:if I was doing it over I'd sort the sentences by kanji to limit the rate of introduction of new readings

Someone wrote a little program that appeared able to sort words based on all the kanji in word, or at least tried to. I had hoped that program could then be used to sort words based on the KO2001 order.

I have this strange feeling of deja-vu...

brandon7s Wrote:I AM still working on memorizing kanji, but since I'm getting so tired of doing nothing but kanji, I figured I'd start on something else to keep it interesting. Also, since I'm not using the actual RTK list, I can simply add whatever kanji I run into while doing Core2000 to my kanji pile. Actually seeing the kanji I'm memorizing being used will help me memorize them better, I believe.

If I was doing RTK over, I'd sort core6k and KO2001 by RTK order, and do the sentences along with the kanji.

Little update on my strategy so far; it's still subject to change though, of course.

So far I have found that using a method similar-ish to TaylorSan's is working best for me at the moment, the main differences being I'm far too lazy to put all the energy and dedication that TaylorSan puts into it.

Oh, and I'm using the standard Core2000 list order, I found the KO2001 order just didn't jive with how I learn - I got mixed up very quickly with all the similar sounding words. After a short while of using that list order, everything became very jumbled up in my head and my memory completely went out the window.

I suspend all the cards, and then I go through and unsuspend about a few sentences. I look up each word in the sentence to see if that vocabulary word is in my deck, if it is, I assign it a HighPriority so that it shows up immediately. If there's a word in the sentence that I DON'T have a vocabulary card for, I make one of my own.

Now, some might consider this next step to be a bad idea, and if anyone has something to say about it, I'd love to hear some opinions. When I make my own vocabulary card for the word that is missing from the sentence, I use Text-to-Speech software (TextAloud+Misaki voice) to create audio for the new card. I do realize that the pronunciation is right maybe 75% of the time (from what I've been told), but since I do have the Core2000 sentence audio that contains the word, I can compare the Misaki voice pronunciation to an actual human being and so far I have not yet run into a Misaki-generated word that sounded wrong, to my ears at least.

So, essentially, my current method is this:
1: Suspend all cards
2: Unsuspend a few sentences at a time, along with any vocabulary cards that are words contained in the sentences.
3: If no vocabulary card exist for a word in the sentences, create card (with Misaki TTS audio)
4: Repeat and, hopefully, PROFIT!

(note: I'm not using any recall cards, just recognition.)

WWWJDIC has an alternative to Text-to-Speech. The Jpod101 people recorded audio on almost all the words, and you can rip them using audacity (or similar program). I have used it to create audio for sentences that don't have any (though now I just skip it, I might check with my japanese friend to see if I'm pronouncing the word correctly). Granted this will be the dictionary form of the word (not a bad thing to learn), but it's not hard to alter it and add particles when you read it out loud. And sometimes they have full expressions and such. Do you use an example sentence on the cards for these words? I don't think it should be "AJATT dogma" that you do, but it if you study new words on separate cards I'd think including them at least in the answer column would be a plus. Actually, I may follow you on this. I've been learning all the words in a sentence, as is, but breaking it up and focusing on each new word might be a real benefit. It can be a bit tiresome learning several new words in one sentence, your method reenforces these words, and the original sentence. I think I would look up the word and use an example sentence that I can read as is.......

Haha - sorry I was really coming up with all that as I typed it.... thanks for the inspiration!

The other thing I'm doing now is making "production" cards for super mature cards, cards that have a due date of 3-6 months. Audio ----> write the the sentence. When they're that mature you don't have to struggle as much to write them cold. Yes it will eventually double my work load, but I'm not a super speed 50+ sentences a day Ankier (not by a long shot), and I think being able to produce the kanji is a great skill that brings it full circle (yes this is debated, but I think it's worth it for me).

TaylorSan Wrote:WWWJDIC has an alternative to Text-to-Speech.

WWWJDIC has audio recordings? If so, I'd love to know where, I haven't been able to find them if they are there..

TaylorSan Wrote:Do you use an example sentence on the cards for these words? I don't think it should be "AJATT dogma" that you do, but it if you study new words on separate cards I'd think including them at least in the answer column would be a plus.

By "these words" I assume you mean the dictionary form of the words that aren't already included in Core2000, but are used in the sentences? I don't think I'll do that, at least not yet. If I start adding more example sentences, it's going to be a vicious cycle of adding words from sentences and then more words from those sentences and then even MORE words from THOSE sentences. At least that's what would happen with me. - when I get good enough to understand most simple example sentences, than that won't be an issue, but I'm at a very basic level right now. In the future, I'm sure I'll switch things up as the need arises. Having a standard set of material to work with (the sentences that come in the Core2000) gives me something to work towards completing; and I need to feel like I'm actually completing something or I'll get distracted and move on to something else.

TaylorSan Wrote:I've been learning all the words in a sentence, as is, but breaking it up and focusing on each new word might be a real benefit. It can be a bit tiresome learning several new words in one sentence, your method reenforces these words, and the original sentence.

The way I look at it: the sentences cards are there for grammar practice, pronunciation, and just to help me develop an ear for Japanese in general. The vocabulary cards are there to teach and help me remember words.
The audio portion of the vocabulary cards are essentially there only to help me memorize the words without seeing them (I find it MUCH easier to remember a word by reading than by listening, so the audio is very important to me). The audio isn't there so much as to teach pronunciation as it is for testing my "ear memory", so to speak.

Also, I think pronunciation is much easier learned by hearing a word in context, since a word said in a sentence usually sounds fairly different than the same word said in isolation.

TaylorSan Wrote:The other thing I'm doing now is making "production" cards for super mature cards, cards that have a due date of 3-6 months. Audio ----> write the the sentence.

That almost sounds like the same thing as recall (or recognition, as some call it). When I review a Listening card I count the review as correct (3, on anki) if I know what each word means, and how it's written. Actually, I tend to think about how a word is WRITTEN first, and then remember what it means after that. The way a word is read/written and it's meaning are completely inseparably, in my mind. The way a word sounds is more of an after-thought for me. I guess this might be the case for me because I am a VERY visual learner.

Ok - Thanks for the responses. I haven't used the priority feature, and I missed that part of your explanation the first time, so I have a better idea of your process now. I have not thought about the process the same way you do, but I think you have some interesting insights. I have approached it by learning sentences, and then studying the individual words from there.

To get the audio on WWWJDIC, click the little blue arrow to the left of the word. Sounds a little nicer then the robot one you use, but you will only get single words. But as I said, there are some full expressions and things there too.

I'm not much ahead of you, but I am just recently finding that I often can read some, or several of the example sentences, or sometimes am only missing one word. It's a rewarding feeling!

I am also mining Jpod101, using the pdf's and audacity to make the cards for anki (separate deck). I rip out the conversation from the lesson and put it in my ipod too. I find it to be a big help for listening to the the conversation in whole, which I have sentence mined into anki, and also can study/read/listen to the individual sentences in the ipod. Yes, it takes time, but getting the context of the conversation is more fun and beneficial, so worth it. I use smart.fm to increase vocabulary, and I like the word order/frequency thing. So I go back and forth with mining. The jpod is more work, and more fun, and seems to be easier to "stick" because it comes from context, but the auto import deck is efficient and valuable. I just make sure I clear the reviews for both decks first, then add new stuff depending on the time I have each day.

TaylorSan Wrote:To get the audio on WWWJDIC, click the little blue arrow to the left of the word. Sounds a little nicer then the robot one you use, but you will only get single words. But as I said, there are some full expressions and things there too.

D'oh! So simple, yet I completely missed that. How embarrassing, haha. Awesome find though! Those recordings sound much better than the Text2Speech I'm using. The only downside is that I have to record the sound manually instead of saving the clip to disk from the website. Maybe one of the firefox extensions for downloading embedded content will work. I'll have to give it a shot. Either way, the sound quality is worth the extra hassle. Thanks for the heads up, TaylorSan.

TayloSan Wrote:I am also mining Jpod101, using the pdf's and audacity to make the cards for anki (separate deck). I rip out the conversation from the lesson and put it in my ipod too. I find it to be a big help for listening to the the conversation in whole, which I have sentence mined into anki, and also can study/read/listen to the individual sentences in the ipod. Yes, it takes time, but getting the context of the conversation is more fun and beneficial, so worth it.

So you recommend signing up for Jpop101? I've heard a few of their podcast, and liked what I hear, but I haven't heard any opinions on the rest of the material they offer.

I think after I've finished, or gotten at least halfway through, Core2000 I'm going to sentence mine the Full Metal Panic: Fumoffu anime series. I've seen all the episodes many, many times, and recently I ripped the audio from them and have been listening to them while I'm playing video games or doing chores (using iPod).
Listening to something I am already very familiar with is a whole lot more enjoyable for me than the alternative, at this point. I only understand about 5% of what's being said, but because I know the series so well I can still tell what's going on even if I can't understand any of the words.

Yeah I rip it manually too (audacity). I myself enjoy the jpod101, some people around here don't (they don't like Peters dorkeyness or think there is "too much English" - if I recall correctly). I think Jpod is a good place for a beginner to start - And much more fun then smart.fm.

Pro's: i+1 lessons that build on each other, context based conversations from native speakers, fairly good dialog explaining grammar and vocabulary nuances/uses, and some humor and personality. The podcasts are free, and you could do the week free trial and download the accompanying pdf's to your hearts content. Really a TON of mining to be had there. They also have lots of lessons that are about pop culture, and contemporary Japanese things/events etc.

Con's: Mining them can be time consuming. Some of the Pdf's don't copy/paste, so I end up typing the sentences. And you do need to use audacity (or similar) to rip the conversation (unless you pay for a premium thing, but even then I think they have the shows intro on there), and to rip the sentences. It can take some time, but I really find that doing the process helps me learn it too, so it's not as wasteful as it may appear.

My system for Jpod is to listen to the lesson in the ipod, (usually while I'm doing the dishes or something), then latter load the lesson into audacity. I then clip out the conversation, and look through the pdf for stuff I can't read (new vocab/kanji). Then make the cards, and also put a snapshot of the kanji sentence on my desktop, to label the mp3's in itunes (I like to have a big display of the sentence for my touch screen - "album cover art" - for each sentence). It takes some work but I learn the material well, and can study it any time, any where as audio/and or reading. And of course I listen to the ripped conversations to train listening as well.

It may not be the most efficient way to learn... I have no idea. But it works well for me, and jpod is a good mixup from just using contextless smart.fm.

Nukemarine Wrote:

cangy Wrote:if I was doing it over I'd sort the sentences by kanji to limit the rate of introduction of new readings

Someone wrote a little program that appeared able to sort words based on all the kanji in word, or at least tried to. I had hoped that program could then be used to sort words based on the KO2001 order.

However, the sorting program didn't really come about, so the only order we have is the order based off newspaper frequency.

now I'm thinking of redoing rtk (or a ko2001 subset plus dependencies) with Japanese prompts and at the same time doing ko2001 sorted in rtk order (sorting on just target kanji or the whole sentence)

here's a script that can be used to sort sentences by kanji: http://ichi2.net/anki/wiki/ContribFugounashi#kanji-sort

cangy Wrote:now I'm thinking of redoing rtk (or a ko2001 subset plus dependencies) with Japanese prompts and at the same time doing ko2001 sorted in rtk order (sorting on just target kanji or the whole sentence)

here's a script that can be used to sort sentences by kanji: http://ichi2.net/anki/wiki/ContribFugounashi#kanji-sort

How does the plug-in work if I may ask? Is it a selection on the sort feature? If I read the code correctly, then the order of the word is based off the "difficulty" of the kanji in it (well, the kanji further down the line). So, 蛇口 will be placed by 蛇 and not the much easier 口 if using the 2k1KO order.

I'm assuming that if I rename the portion to whatever I call the field with vocabulary words then it'll sort according to that instead.

Now the big question: Can the list be exported as a text file in that order?

The reason I ask is then I can, sort iKnow Core 2000 and 6000, export the list, sort the kana only words back to their original settings, then spread those throughout the 2001KO sorted words. Post new sorted list on Anki.

Still, thanks for the plug-in.

Nukemarine Wrote:How does the plug-in work if I may ask? Is it a selection on the sort feature? If I read the code correctly, then the order of the word is based off the "difficulty" of the kanji in it (well, the kanji further down the line). So, 蛇口 will be placed by 蛇 and not the much easier 口 if using the 2k1KO order.

I'm assuming that if I rename the portion to whatever I call the field with vocabulary words then it'll sort according to that instead.

Now the big question: Can the list be exported as a text file in that order?

The reason I ask is then I can, sort iKnow Core 2000 and 6000, export the list, sort the kana only words back to their original settings, then spread those throughout the 2001KO sorted words. Post new sorted list on Anki.

Still, thanks for the plug-in.

ok, so here's a few things to note. first, it's a stand-alone perl script rather than an anki plugin, so you'll need to have perl installed to use it. also, it sorts text files, so you'd need to export your deck facts or save your spreadsheet to a tab-separated text file. and as you said, it sorts based on all the kanji in the field and the furthest of their positions in the sort list (so "most difficult" in ko2001 order). lines with fields that don't contain any of the target kanji will be sorted to the front

so, say you export your deck as exported.txt and the field you want to sort on (the sentences, say) is the 3rd field (number 2 starting from 0), and you download kanji-ko2001.txt as the sort order, then you can produce the sorted file by running the script like this:

kanji-sort --kanji kanji-ko2001.txt --sentence-field 2 < exported.txt > sorted.txt

you'll also get 2 new fields appended to the sorted output: the index of the kanji in the sort list of the highest kanji used in that line (i.e. which ko2001 kanji you have to be up to to be able to do that sentence), and the increase from the previous line