I've finished RTK, and someone recommended knowing some basic grammar rules before I head off to do some sentence mining. Which makes sense.

So, I'm going through Tae Kim's Complete Guide and he gives you some Vocabulary to learn. Obviously, Im not going to remember them just from reading through it in one go. So, I thought I should use this chance to make a Vocab deck on my SRS.

Only, I'm not sure what would be considered the right way to do it. Or rather, an efficient way. What side do you put the information?

Do you do something like..

FRONT:日本

BACK: 【に・ほん】 - Japan


Or..

Front: 日本 【に・ほん】

Back: Japan

Or is it something completely different?

If I actually did a word deck, which I don't, it'd be the former.

FRONT:
日本

BACK:
にほん
Japan

--
I do sentences though so my cards look like.

FRONT:
彼女は観客を魅惑する演技をした

BACK:
She gave a performance that enchanted the audience.

enchantment; fascination

みわく

彼女 【かのじょ】 (pn) (1) she; her; (adj-no) (2) her; (3) girl friend; girlfriend; sweetheart; (P); ED
観客 【かんきゃく】 (n) audience; spectator; spectators; (P); ED
魅惑 【みわく】 (n,adj-no,vs) attraction; fascination; lure; captivation; charm; (P); ED
演技 【えんぎ】 (n,vs,adj-no) acting; performance; (P); ED

I'd say it depends on where you're at. For you first 2,000 basic words, it's good to go for a production method using something like photos (best), translation combined w/ cloze delete (has problems), kanji to kana and kana to kanji cards (hits both directions, but has problems).

After your basics, you can go with recognition or continue with the more difficult production methods. For sentence cards such as from subs2srs or figuring out a novel, then recognition should be fine.

Reason I recommend production for the basic words is these words will be used a lot. Being able to bring them to the forefront of your mind on call is a great thing and will help you talk or write around weaknesses in the less common words. Now, I don't know where you can get a good set of photo production cards. There might be a streamlined Rosetta Stone card set someone put out there, but I haven't seen it. If you're cool with cloze delete or recognition cards then you're set as many deck types exist for those.

@vix86
Ah. In other works, its probably not best to learn vocabulary as a stand alone. Tae Kim mentions I should learn it in context, and simply gives you a list. And then at the end, declares that you know it all. Which doesn't make much sense to me.
I cant really tell whether he just wants to expose me to it or if he wants me to learn it, I have half a mind to just drudge through the Grammar section with just my knowledge of RTK. Just to get a feel for how sentences work, so I could attempt to do something along the lines of what you're doing.

@Nukemarine
I don't think I'd mind cloze delete. For the most part, I'm just trying to find a way into actually studying the Japanese language. Everyone seems to have their own sentence deck or vocabulary deck or whatever. Since I'm just starting out and there is no clear path.

The only way I can think of is, I should start with a basic sentence and then completely decipher it and while doing that, it'll introduce a new sentence or a word that I could use to find example sentences for, then Ill decipher those, ect.

@CloverJoker, I started working through Tae Kim right after finishing RTK as well and this is what I did.

I went through a lesson each day and added all the vocab words that I didn't know to a vocab deck using Rikaisama. After that I went back to the beginning of the lesson and read through it (read through each example sentences to make sure I understood it completely). Then I opened up Nukemarine's cloze delete Tae Kim deck (available on Anki shared decks) and tested myself on the sentences from the section I just finished reading. If I got any wrong I would go back and read that section again to make sure everything is clear.

I ran through my vocab deck whenever I had time since most of the words stuck pretty easily after all that other work.

Okay, thanks to you guys, I think I've come up with something..Would this be okay?

FRONT- あの人誰？


BACK:Who is that man?

【あの】-say; well; errr ...; (P)

人-【ひと】
(n) man; person; human being; mankind; people; character; personality; man of talent; true man; another person; other people; others; adult; (P)

誰- 【だれ】
(pn,adj-no) who; (P)

I think that works fine as a layout, but in this case あの means "that."

Ohhhh...okay. That makes a whole lot more sense.

Anyway, thanks for all the help guys. I was a little confused about this process before and was unsure how to progress and was stuck, but I now I think I got it.
Now that I have this sentence, I think I can just repeat the steps to make more sentences similar to this one along with Tae Kim.

Seriously, thanks!

CloverJoker Wrote:Okay, thanks to you guys, I think I've come up with something..Would this be okay?

FRONT- あの人誰？


BACK:Who is that man?

【あの】-say; well; errr ...; (P)

人-【ひと】
(n) man; person; human being; mankind; people; character; personality; man of talent; true man; another person; other people; others; adult; (P)

誰- 【だれ】
(pn,adj-no) who; (P)

Thats fine.

Also the bottom part on mine, which you seem to be using can be generated using a plugin on Anki called "Glossing" (or something like that. I can't recall exactly and can't check right now.) It uses WWWJDIC's text glossing and can generate glosses for 100's of cards in minutes.

vix86 Wrote:Thats fine.

Also the bottom part on mine, which you seem to be using can be generated using a plugin on Anki called "Glossing" (or something like that. I can't recall exactly and can't check right now.) It uses WWWJDIC's text glossing and can generate glosses for 100's of cards in minutes.

Oh wow, that's fast. That beats copy and pasting.
Seems to generate more stuff than I need, I take it you go back and delete the unnecessary stuff. Like..for 人.

人 【じん; と; にん; ひと】 (じん) (n) man; person; people; (と) (suf) person; (にん) ; (n) man; person; human being; mankind; people; character; personality; true man; man of talent; adult; other people; messenger; visitor; (ひと) ; (n) man; person; human being; mankind; people; character; personality; true man; man of talent; adult; other people; messenger; visitor; SP

And for this sentence, I would shorten it down to..

人 【ひと】 (n) man; person; human being; mankind; people; character; personality; true man; man of talent; adult; other people; messenger; visitor; SP

Yes trimming the extra out will have to be done by hand. Though sometimes I just leave it for quite a while. My Gloss field is a pretty small font in my decks comparatively. Most stuff is like 40-50 point font while my gloss field is 12 point. The gloss field is only there if I need it. It eliminates the need for a second sentence with furigana for words + the need to check definition if I forget a word.

Don't forget that it's just as easy to copy in simple japanese dictionary definitions. I don't think there's an automated way to do it though.

Maybe someone can try to incorporate the Tanuki definitions which a pretty simple Japanese only definitions.

I've been working with this format for quite a long time now and it works best for me. Here's what I do.

FRONT:

あの........誰？

Who's that man?

--------------------------

BACK:

人

ひと

あの人誰？

So I have the cloze sentence and hint on the front and the expression, reading and example on the back. When you get more advanced, you'll probably want to chage the language of the hint to Japanese.

Of course, with such simple words as "man", I'd probably just write "man" on the front and "人《ひと》" on the back, but this format is perfect for more abstract, vague words or for cases when you have many similar words and you need the context to help you choose. When a word has more than one meaning, I make several cards with different cloze sentences.

Thanks to you guys, I think I've come up with a decent way to do it.


Is it alright that I'm missing the cards a lot because of the readings? I seem to be doing okay, as far as knowing what the sentence is about but I miss the same thing over and over again because I can't remember the reading. Although, eventually I do get it right, which is fine with me and since I just started, it seems kind of normal.

But you know what the say about taking that first step..

Try reading the sentence out loud.

Also -- you don't necessarily have to know all the readings for all the words that Tae Kim introduces right away. If you're having a lot of trouble with a particular reading, I don't think it would be a bad idea to put the reading on the front of your card, like so:

あの人誰（だれ）？

You don't want to introduce too many unknown elements with each new card.

CloverJoker Wrote:Is it alright that I'm missing the cards a lot because of the readings? I seem to be doing okay, as far as knowing what the sentence is about but I miss the same thing over and over again because I can't remember the reading. Although, eventually I do get it right, which is fine with me and since I just started, it seems kind of normal.

I would say this is probably alright. I've been studying Japanese for a while and I've gotten to the point where I seldom miss a reading for a card. It's generally much easier to remember than the meaning.

But when you're first starting out, it's pretty hard to remember readings. Just keep trying and it should get easier.

Also, it might help you to make a some "reading" cards, where all you get is the Kanji on the front and the kana on the back. This should be easier and faster than having to remember the meaning as well as the reading.

I personally do it this way:

人
person, people
-----------------------------
ひと

There is after all some kind of link between the meaning and the kanji writing (when kanji is used at all). To me this is the most logical way to do since in real life, I want to 1. be able to read books aloud, so to speak 2. understand the meaning of spoken sentences. The only little problem being homophones, but adding a hint like "nature" or "body" to distinguish 葉 and 歯 is enough. You can adjust the settings in Anki so that you can't have twice the same field. This way, you won't ever have unnoticed "doubles".

Again, just the way I do it, but I don't incorporate sentences. It's fine to want to learn to use the right word in the right context straightaway, but I prefer getting the building blocks together first.

Just a note. Doing it that way, you never test whether you are remembering the meaning of the kanji/word as well since you are sticking the meaning on the front of the card. You are simply assuming some link is being formed by having the two together and that may not be the case.

The cards are forward & reverse. When I'm presented with the kanji + meaning, I've got to find the pronunciation. When presented with the pronunciation, I need to find the meaning and the way it's written.

You know what would be cool? A standard picture deck for the 3-6,000 basic words/concepts you should know for any language. (For production.)

Picture on the front, leave the back blank (maybe the English word so we know what's being tested), so you could otherwise modify it for any language you learn.

Having one standard deck like that would save a bunch of time.

rich_f Wrote:You know what would be cool? A standard picture deck for the 3-6,000 basic words/concepts you should know for any language. (For production.)

Picture on the front, leave the back blank (maybe the English word so we know what's being tested), so you could otherwise modify it for any language you learn.

Having one standard deck like that would save a bunch of time.

This guy suggests a similar vocabulary concept and gives about 400 words in English that are good to learn in any language.

On your idea, is there a way to have a group project where a list of words are populated by photos, then equivalent in other languages? Probably already exists though.

I was thinking along the lines of the post a couple of weeks ago about the giant list of words from kuro5hin that everyone should know in a language... I think that was around 6,500 words or so?

I would think you just need to grab that list (or a similar list), gather the interested people, and break it up into chunks. Upload collections to Google Drive or DropBox, or wherever you want to store them.

Images have to be small for this to work. 6,500 images at 100k gets in the neighborhood of 650MB (not exactly, but close enough.) at 25k per image, you can cut that down to 162.5 MB or so. (50k would even get you down to 325MB.) You'd need something that can take photos and resample them at "thriftier" data levels. (Like Photoshop, or maybe something that can do batches?) They won't look as good, but it's not an art contest.

Even better, if there was a script that did all this for us. (Or if it already exists?) Why reinvent the wheel, right?

rich_f Wrote:You'd need something that can take photos and resample them at "thriftier" data levels. (Like Photoshop, or maybe something that can do batches?) They won't look as good, but it's not an art contest.

Even better, if there was a script that did all this for us. (Or if it already exists?) Why reinvent the wheel, right?

Photoshop can do batch conversion, but it's not an appropriate tool to do what you're saying (it's like "killing an ant with a nuclear weapon"); it's better to use something lighter and free such as XnView or Irfanview, which have actually a better suited (and much easier to use) batch conversion and resampling support. And, if you want a script, there's ImageMagick for that.

Hyperborea Wrote:It might be better if you didn't store the pictures directly. That could lead to size issues, as is mentioned above, but also copyright. It might be better to have links in the spreadsheet to images in Wikimedia Commons.[...]

That could work, but what if people want all the images to have the same maximum height (so that it fits better in Anki's review window) and have them all ready for offline use? Hotlinks would be worse for those many people; downloading a package with all the images ready to use would be better, and there are many places that can easily store a 325 MB file size mentioned by rich_f, so that wouldn't be much of a problem. As for copyright, it's perfectly ok to package the images from the Wikimedia Commons website you posted (some need attribution, but that can be accomplished by including said spreadsheet in the package). Flickr also has many public-domain and CC images that can be used for this, and for other images maybe there's also Fair Use depending on how this material is distributed.

Yeah, I was thinking of CC photos as well. Or just grab a camera and do it yourself for either really easy thing, or things you have trouble "getting." If you need a photo of a pen, why download something if a photo of your favorite pen works better? And if you're having trouble remembering a particular word, what better way to drill it in than mumbling it while you're trying to capture it in your camera? (Ok, you'll look creepy if you do that but...)

A lot of these basic words are ... well ... basic. It's the same thing with RTK stories. The ones that are the most personal tend to be the most effective. I found that copying someone else's story wasn't as effective as using my own when I did RTK.

And yeah, I have Photoshop 5, and it will probably tear though this kind of task, but Photoshop is probably too much for something this simple.

As for storage, Google Drive promises ~5GB, Microsoft has more, and Dropbox has 2GB. If you're willing to roll your own fileserver through dyndns or something similar, you can save yourself the trouble entirely.

EDIT: Here's the post about the list:
http://forum.koohii.com/showthread.php?tid=9231

It's around 2500 words... so for some of us, it won't be as useful for Japanese (core gets you 6k), but it could be useful for other languages. Also, looking at the words, a lot of them *do* look difficult to visualize. (Just looking near the top, a lot of words like "ability/capacity", and "accustom" look tricky to convey with the proper images.)

Might be easier to do with 1000 or so really basic words, now that I think about it.