+- kanji koohii FORUM (http://forum.koohii.com)
+-- Forum: Learning Japanese (http://forum.koohii.com/forum-4.html)
+--- Forum: General discussion (http://forum.koohii.com/forum-8.html)
+--- Thread: Re-examining your SRS workflow (/thread-3447.html)
Re-examining your SRS workflow - vgambit - 2009-07-15
I didn't mean to make it seem as if I was offended by your post, because I wasn't. I was just saying that normally, I would try to do as many as I can, not that I had intentionally done the "average" amount. That is, I usually wouldn't do so few at a time, but I've been pretty busy at work lately. I work a 5 hour shift, and I was just unable to add any kanji for about an hour due to work tasks.
But yeah, it seemed pretty effective, although since I waited until the next day to review (just about 24 hours), I ended up remembering roughly half of the kanji I added. Once I get used to the new routine, though, I think I'll start doing better (and part of it might be the fact that I just switched to RTK Lite as of frame 894, so I'm getting a lot less repetition).
Re-examining your SRS workflow - blackmacros - 2009-07-15
Ah okay. No harm, no foul then
Let us know how it goes. The only concern I have about SRS'ing this way is that for people with actual commitments (unlike myself
), because it lets you add and review many more cards per day, it might be easy to accidentally get yourself into a position where reviews become excessive and missing a day could result in a huge backlog of reviews.
Re-examining your SRS workflow - vgambit - 2009-07-15
Being "too productive" is a downside? lol. Right now I'm at 299 pending reviews + 15 new cards, so I'm in trouble as it is. This makes it a lot easier to whittle down, though (as long as I can find time to keep at it).
Re-examining your SRS workflow - skeeter3 - 2009-07-19
Ah! Just wanted to drop a quick message that this "new method" worked wonders for me today in my trial run. I was able to double the number of kanji(still doing RTK) learned. I managed 92 kanji in the time I studied and had a clear review pile as well. I plan to use this all next week and have all the kanji done finally! Thank you for taking the time to post this seemingly simple method. I will update any with any changes in progress/results that I see as this week goes on.
Re-examining your SRS workflow - RobotsAnger - 2009-07-19
Just an update, I tried this method and it worked great for the first week I was able to pretty much double the amount of items added a day, but for the next week the reviews were too much and I ended up not adding any kanji at all. Working 9-5 didn't help either. So in summary this method seems very be very effective but mostly for those who have the time.
Re-examining your SRS workflow - blackmacros - 2009-07-20
skeeter3 Wrote:Ah! Just wanted to drop a quick message that this "new method" worked wonders for me today in my trial run. I was able to double the number of kanji(still doing RTK) learned. I managed 92 kanji in the time I studied and had a clear review pile as well. I plan to use this all next week and have all the kanji done finally! Thank you for taking the time to post this seemingly simple method. I will update any with any changes in progress/results that I see as this week goes on.Good to hear its working out for you, keep us updated
RobotsAnger Wrote:Just an update, I tried this method and it worked great for the first week I was able to pretty much double the amount of items added a day, but for the next week the reviews were too much and I ended up not adding any kanji at all. Working 9-5 didn't help either. So in summary this method seems very be very effective but mostly for those who have the time.I was afraid that this might be happen (see my post 3 posts up).
I think, though, that you might be able to lessen the damage by adjusting the review/add ratio. For example, you said you were quite busy with your job and started getting swamped with reviews. In situations like that you could adjust it so that instead of review 30, add 15 (a ratio of 2:1) you could start using a ratio closer to 3:1 or 4:1. Add 5 cards and review 20, for example. You're still adding cards (and probably still progressing faster than if you were single-focus SRS'ing) but by adjusting the review/add ratio you can self-correct as it were, by reviewing more and adding less (and therefore generating less future reviews) until you recover.
Re-examining your SRS workflow - mafried - 2009-07-20
As blackmacros said, just don't add as many cards. I've been using the new techniques since this thread was started, and it's cut down my review time considerably. I haven't really been adding any more cards though since, like you, I'm working full time and have lots of other commitments.
It's nice that reviews don't take up as much time as they used to though.
Re-examining your SRS workflow - AmberUK - 2009-07-20
I think this is what makes RTK hard. As its such a large number of cards/data its easy to keep adding. And then a week or month later it bites you on the ass.
I found with lessons in books that doing the vocab first and then the grammar ment there was a break in input meaning there was less of a bite at that later date.
Re-examining your SRS workflow - skeeter3 - 2009-07-26
This method of switching back and forth worked great for me this week. I made a few modifications to suit my learning style as the week went on. Instead of doing a set number of cards for each review I would do only a half hour of review before switching back over to learning cards(set anki time limit instead of card limit). When I switched to learning cards I tried to break the lessons into chunks of around a half hour as well. By leaving lesson goals in there I found I was often more motivated to complete the new kanji. Don't know how long I could have maintained the 100+ kanji/day, but it got RTK finished like I'd planned. Now, if you'll excuse me, I need to go tackle the monster review pile that awaits.
Re-examining your SRS workflow - rich_f - 2009-08-04
Since this thread has been dug up again, I'll ask:
Has anyone tried mixing this with using Iversen's method to getting vocab down the first few times? (He says as he stares at a huge pile of data he needs to add to his deck relatively soon...)
Re-examining your SRS workflow - vosmiura - 2009-08-04
Some others mentioned this, but what I found useful is to have a timer that I set to count down 10 minutes and during that 10 minutes focus on one thing. In my case I switch between reviewing (sentences or rtk), or creating new sentences, or memorizing new vocab from my new sentences. I think doing these small units of time make it easy to stay focused and get through a lot.
rich_f: That's exactly what I do - one of my 'tasks' so to speak is memorizing any new vocab in a similar way to Iversen's method. However I'm not one of the "100 new cards per day" people, I'm one of the "short time per day allocated to studying" people, so I just try to make good use of that time.
Re-examining your SRS workflow - rich_f - 2009-08-04
vos: Thanks for the suggestion. The trick for me is trying to figure out how to adapt Iversen's method to how I study. The whole going from language X to language Y to language X again is tricky when you start adding kanji and kana readings.... and since I do a lot of production, I'm trying to find a way to make it all work.
I guess I should stop worrying and just starting doing it, and change it to suit my needs as I progress.
Re-examining your SRS workflow - vosmiura - 2009-08-04
I think of it as just a way to initially get the words in my head. I have my list as kanji-kana-English and I first memorize going from kanji to kana & English, and after English to kana. But after I've done this process I don't do it again - I just do my SRS sentences and let the meaning cement itself from context.
Re-examining your SRS workflow - ghinzdra - 2009-08-04
This principle as all good principle is not something new....
for those of you who are in detective novel you should know that dr watson said something very close to that about the way sherlock holmes worked... Unfortunately I can t quote him exactly but it must be something like "sherlock holmes didn rest whenever he felt bored he just changed his work"
otherwise
1 kiddo to your achievements....
I was wondering if we should nt try to write a hall of fame/hall of heroes with the name and achievements of everyone. I think it would give a boost to everyone.
2 categories : help to the community (heisig for his book , fabrice for this website , revok for Anki , cb4960 for sub2srs,etc... )
personal achievements : completion of the RTK KObook within X days/month , passed the JPLT level 2 or 1 within x months , first to do this and that and so on
2 storming through RTK, KO2001 ,etc... is a always a good thing whatever the critics may be
It gives you a momentum,confidence that you can use after that to tackle real japanese .
what s more the shorter the chore , the higher your interest
3 I still think that your pace is something very specific to this stage though: you have some material already available thanks the effort of other people.... the only thing you got to do is review . It s very different from the next step where you take your stuff from the environment : picking up sentences takes time especially since it s not formated .
4 the achievement should be clearly assessed : it s extremely unlikely that the audio recognition skill pars with the reading skill ... even with the Iknow file / misaki (and as I used misaki throughout my KO study I have a good idea of the maximal befenit you can get from this) there is no shortcut for audio recognition : you just got to hear a lot of japanese to get accostumed to accent , flow,etc....
5 i m a a bit surprised by the statement about haruhi: i ve completed KO2001 for a while now and I read haruhi and I can ensure you that while it gives you a decent command on it you ve got a new word every two sentences on average . At least.
Maybe you re thinking compounds and onomatopea (like how kyon describe the way mikuru is crying there must be at least a dozen different onomatopea... ) aren t new vocabulary ....
Re-examining your SRS workflow - nest0r - 2009-08-04
ghinzdra Wrote:This principle as all good principle is not something new....Sherlock Holmes also injected cocaine when bored and played the violin when he was thinking. ;p
for those of you who are in detective novel you should know that dr watson said something very close to that about the way sherlock holmes worked... Unfortunately I can t quote him exactly but it must be something like "sherlock holmes didn rest whenever he felt bored he just changed his work" ...
Re-examining your SRS workflow - blackmacros - 2009-08-04
IceCream Wrote:@blackmacros:I feel *awesome* after KO haha. Reviews have dropped to about 450 a day (down from peaks of ~800).
How do you feel now that you finished KO2001? How are the reviews on all the stuff you've done working out? Are they getting less heavy yet? And, are you now entering stuff from japanese you just come across? what do you feel the differences are with the sentences you enter now? what areas are you finding easiest and which are the hardest for you?
I was mining Chrono Trigger (with the help of a really useful spreadsheet, containing all the dialogue (http://www.chronocompendium.com/Term/Retranslation.html) but I got bored of playing it so stopped.
Right now I'm reading Full Metal Alchemist without a dictionary (and not mining it) for enjoyment, and get about 70% comprehension. I'm also mining the Haruhi Lightnovel. I look up every new word I don't know (which is quite a few, probably about 2 or 3 a sentene as ghinzdra was saying). I'm using dictionary sentences for that though, because I still can't fully understand the Haruhi sentences. The plan is to mine it, then go back a few months from now and be able to read it for enjoyment.
I'm also going through the KM2kyuu grammar book at a reasonable pace of 30 sentences/day and after that will go through the KM1kyuu book in preparation for sitting JLPT1 in December (which will be about 9 months since I started studying Japanese...not quite the fabled 3 months but oh well )
The hardest areas for me are grammar and slang, because they aren't as easy to look up as new Vocab.
ghinzdra Wrote:This principle as all good principle is not something new....That's awesome! I didn't know about that.
for those of you who are in detective novel you should know that dr watson said something very close to that about the way sherlock holmes worked... Unfortunately I can t quote him exactly but it must be something like "sherlock holmes didn rest whenever he felt bored he just changed his work"
Quote:5 i m a a bit surprised by the statement about haruhi: i ve completed KO2001 for a while now and I read haruhi and I can ensure you that while it gives you a decent command on it you ve got a new word every two sentences on average . At least.I'm not sure what you're referring to here. Was somebody talking about Haruhi previously and I missed it?
Maybe you re thinking compounds and onomatopea (like how kyon describe the way mikuru is crying there must be at least a dozen different onomatopea... ) aren t new vocabulary ....
EDIT: Or was it me talking about Haruhi previously and I've forgotten. If so I certainly don't think its easy or anything, in fact even after mining it I don't grasp it very well. The sentences are tough and long
. 2 to 3 new words a sentences sounds about right.
Re-examining your SRS workflow - mafried - 2009-08-04
rich_f Wrote:Has anyone tried mixing this with using Iversen's method to getting vocab down the first few times? (He says as he stares at a huge pile of data he needs to add to his deck relatively soon...)Yes. It works quite well, with modifications.
I'll write more when I get home, but I image vosmiura and I are doing something similar.
Re-examining your SRS workflow - mafried - 2009-08-04
So the Iverson's method is big on translating base language -> target language. I don't think I need to explain to anyone here the problems in doing this. In fact, the only part of "Iverson's method" I recommend at all is keeping a list of vocab you have trouble with, and learning in blocks of 5-7. In fact, in retrospect it's a pretty terrible method that just happens to have those two real gems embedded in it. But that is what I assumed you meant, in which case yes, "Iverson's method" has worked for me. Here's how:
During blocks of time devoted to "adding new cards" I do the following for each new sentence fact in anki:
1. Look up all grammar and vocabulary I don't understand.
2. Add vocabulary is to a running list.
3. Unsuspend card.
4. (Only) if list has reached 6 or more items, learn the first six vocab in bulk. I use mnemonic techniques mostly, and don't bother writing out vocab like Iverson recommends.
5. Move on to next fact.
When reviewing, I also add vocab I failed to remember to that list, but I wait to review those items until I move to an "adding new cards" chunk of time, even if the list grows past six vocab.
One more note: I don't bother with the multiple meanings of a vocab. I just learn what it means in the context of the sentence that I found it. The other meanings will come with time, and this simplifies things greatly.
I've been doing this off and on--when I really have to because the vocab load is overwhelming--for a long time (almost a decade with various languages) and it really works. Only recently I've systematized it and started learning ALL vocabulary in this manor, and that's taken it to a whole new level. I'd say my efficiency while learning/adding cards has increased to 2x or 3x what it was before.
EDIT: I should add the implementation details which I know someone will undoubtedly ask about. I keep the list as kanji - kana - english, where the english is the direct literal translation of that word for the sentence (if you've made the switch to J-J monolingual, I assume you could put a simple Japanese definition or notes or something instead). My "review" of this list is testing myself left-to-right (i.e, kanji->kana->english). Once I am confident that I have learned it I move on. When I finished all six I cover the kana and English columns, then close my eyes and for at least 15 seconds do/think about something else. Then I open my eyes again and, starting from the top, read down the kanji column trying to recall the conceptual meaning (and kanji readings) of each word. If I'm successful, I cross it off the list. If I fail, it gets added back onto the list for next time.
Also, I alternate between adding cards and reviewing the list. Even if I have 12 or more vocab on the list, I'll learn a card, review six, learn a card, review six, etc. until done. I do it because I find both activities hideously boring and it makes it the whole thing manageable. But maybe I'm mentioning it because maybe it does make a difference for you too.
Re-examining your SRS workflow - ghinzdra - 2009-08-04
blackmacros Wrote:If I can give you a piece of advice if you don t get it/Understand then don t get it/mine it .ghinzdra Wrote:This principle as all good principle is not something new....That's awesome! I didn't know about that.
for those of you who are in detective novel you should know that dr watson said something very close to that about the way sherlock holmes worked... Unfortunately I can t quote him exactly but it must be something like "sherlock holmes didn rest whenever he felt bored he just changed his work"
Quote:5 i m a a bit surprised by the statement about haruhi: i ve completed KO2001 for a while now and I read haruhi and I can ensure you that while it gives you a decent command on it you ve got a new word every two sentences on average . At least.I'm not sure what you're referring to here. Was somebody talking about Haruhi previously and I missed it?
Maybe you re thinking compounds and onomatopea (like how kyon describe the way mikuru is crying there must be at least a dozen different onomatopea... ) aren t new vocabulary ....
EDIT: Or was it me talking about Haruhi previously and I've forgotten. If so I certainly don't think its easy or anything, in fact even after mining it I don't grasp it very well. The sentences are tough and long
after months of experiments I have come up with 3 SRS laws that are not to broken on any circumstances as far as I m concerned
a good SRS fact/card should be
1st law UNDERSTOOD in every part , every word
as I said earlier if you don t get it(Understand) then don t get it(mine it )
If you really want to learn it / you have the feeling that you ll eventually get a grasp of it then suspend it with a tag like "difficult" and after a month take a look at it again
2nd law SHORT : ]no more than a sentence/ 10 sec
except in case of conflict with the First Law.
if it s longer and I still crave for it
for the written source I make as much copy of the fact as needed and I underline a different part in each copy . This way I focus on a short part but I always have the background if needed .
for the audio/sub2srs source I split the sentence in as much part as needed and I keep a copy of the long sentence suspended . When i feel really at ease with all those part I put the long sentence back in the revision pile and I suspend the short ones.
3rd law USEFUL/INTERESTING
as long as it does nt conflict with first and secund law
as for me that disqualifies almost totally haruhi as mining material. It is interesting but is wayyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy too long .At least for the Kyon reflexion.
Re-examining your SRS workflow - blackmacros - 2009-08-04
ghinzdra Wrote:If I can give you a piece of advice if you don t get it/Understand then don t get it/mine it .Yeah I totally agree, in that if you don't understand something it should not be in your SRS. Something like Haruhi is still too much above my level, but its something I'm interested in and want to read someday. So I go through and 'mine it' but I use dictionary example sentences for each word instead of Haruhi sentences, to make sure I actually understand what I'm putting in my SRS. Whereas with other stuff like Chrono Trigger and FMA I can often understand it quite well, so I use the source sentences.
Re-examining your SRS workflow - ghinzdra - 2009-08-04
did you find an extractor to get the script of chrono trigger out of the DS?
I know there are softwares on the net but it seems like it s a bit complex. Either you have a general software that doesnt work 9 out of 10 times or you need an extractor specially built for the soft
I d love to extract the script of the 逆転裁判 games.... I know there is one for the GBA version..
Re-examining your SRS workflow - blackmacros - 2009-08-04
This site has the full Japanese script, as well as the English game translation and a more accurate fan retranslation. It's awesome, made mining it very easy.
http://www.chronocompendium.com/Term/Retranslation.html
Re-examining your SRS workflow - ghinzdra - 2009-08-04
Wait are you aware that the DS edition is special ? they have changed a lot of things with brand new quest , pokemon-like collection of monster and a lot of stuff like that....
Re-examining your SRS workflow - blackmacros - 2009-08-05
No I wasn't actually. I was just playing it on a SNES emulator, which is what the script is for. I'm not sure how extensive the changes are in the DS version, so maybe it wouldn't be as useful.
Re-examining your SRS workflow - skeeter3 - 2009-08-05
Blackmacros, being a jobless college student as well, I am looking at adding a little KO2001 to my study during this last month of summer before school starts again. I posted earlier with the successful results I got from using this method with RTK. My problem with KO2001 is the large vocab wall. I opened the book, looked at the first few sentences, and was floored. I would like your, and others', input on how you dealt with the steep curve of the vocab please. How did you remember all that vocab, especially when you're adding ~200 sentences per day? Thanks!