Do you think iKnow is worth subscribing to?

What do you think?

EDIT: Sorry about (probably) posting to the wrong forum. If some moderator has time, please move this thread to the Language resources forum.

Don't they have a free trial?

I just want to know what other people think about it. What does iKnow offer that I can't get from using Anki + Core 2000 and 6000 decks?

I've tried quite a few different sites and programs, including iKnow (before, during and after it turning into Smart.fm), Renshuu.org, ReadTheKanji.com, Anki, and more.

iKnow is the only one I've found that repeats the same word over and over during the study sessions. It'll do it from JP->EN, EN->JP, Kanji->Typing, Audio->EN, Audio->JP, Audio->Typing, Audio Sentence->Typing... I think that's the full list. This repetition is the best way for me to learn a new word.

All the others work basically like Anki. If you get it wrong, it shows up again soon, but if you get it right, it goes away for a while. Sometimes a long while.

The Core word lists are pretty decent, but other more optimal lists have been created in the past. Kore, etc.

Edit: As a programmer, I keep thinking I should create a better site than iKnow, but usually I end up thinking that there isn't much I can improve on it. Especially now that it has custom word lists again. The only thing I can think of would be an API, which they used to have, but don't now... And don't plan to do again. -sigh-

I got to step 3 of core 2k anki deck and it just shut me down. Of course I was trying to know the entire sentence. I was told later that this is not the type of deck to be doing that with. So I tried out the iknow.jp website. I like the new word order much much better! I think I started over around step 5 of the core 1000. (Not too many repeat words from the old step 3 anki deck)
Its cut down on my study time some. I really hate and love the way it makes me type the word out. At first I couldn't really do it well. Now I can hear most words and type them quickly.

I guess its up to you. I enjoy it.

Go ahead if you are made of money. US$150/yr is crazy when the decks are on anki already for free. They have a long history with bugs in the system and ignoring users complaining about them as well.

Superfreek Wrote:I think I started over around step 5 of the core 1000. (Not too many repeat words from the old step 3 anki deck)

You can uncheck the words you already know and then you don't have to re-study them. That's what I've been doing. That's another reason I like iKnow, actually. Most sites make that tedious or impossible.

Money isn't really the problem. I am not remotely close to rich or anything but I don't think $150 for 1 year is "crazy". However, I don't like wasting money either.

Anki has all the sentences already and it's free. iKnow presents a nice online interface to review them, but is it really worth it? Anki is winning so far.

I think they have 5 free study lessons. Try it out. If you like it pay, if not anki is not bad either. I just like the new word order on iknow.jp. Maybe someone has already has an anki deck in the new order. If so I would try to find that one.

Good luck.

I found iKnow really good for learning Japanese vocab. For me, Anki is good for retention but not initial learning.

Before it became a paysite I used to do each lesson to 100%, then import that lesson to anki for long term study. The way it does kanji recognition, and listening, and kana > english meant i could always actively recall the word, and also the sentences were good. At first i couldn't understand them, but by about word 500 they were all starting to make sense.

However, I don't find it so useful now I'm studying Chinese. Their beginner Chinese course just teaches 500 characters, 1 at a time - no compounds! So you end up studying fragments of words, which is stupid - just teach us the whole word!

But if you have the money, give it a go, study a few times and see if it works for you.

Slime Forest Adventure is $25. I have a key for two free copies, so please PM me if you want one (note that this means it's $0 for you). It is really only good for testing kanji->keyword. Heisig says not to do this; see other threads for more info (search on kanji and keyword in thread title). Some attempt has been made at testing other things, but it's not very good. Anyhow, it has an adaptive algorithm that will ask you questions you get wrong more often, completely ignoring the "leech" theory (like iKnow! does).

jcdietz03 Wrote:Slime Forest Adventure is $25. I have a key for two free copies, so please PM me if you want one (note that this means it's $0 for you). It is really only good for testing kanji->keyword. Heisig says not to do this; see other threads for more info (search on kanji and keyword in thread title). Some attempt has been made at testing other things, but it's not very good. Anyhow, it has an adaptive algorithm that will ask you questions you get wrong more often, completely ignoring the "leech" theory (like iKnow! does).

I learned my kana with it and it was amazing for that.

I also did 1300 kanji on it, but only the easiest ones stuck for any amount of time.

You might already know this, but you can configure your Anki deck to force you to type in the answer. If you want tons of repetition, you could even make multiple versions of the same fact by changing what goes into the expression/answer fields for cards.

jishera Wrote:If you want tons of repetition, you could even make multiple versions of the same fact by changing what goes into the expression/answer fields for cards.

True, but those cards aren't guaranteed to stay together, and failing 1 of them doesn't fail the rest.

Anki does what it set out to do really well: Retention.

It just doesn't do -learning- well enough for me.

1. Anki is customizable enough to allow you to emulate or even improve on the iKnow format.
2. Learning to do this, then actually doing it/revising it, could take quite a bit of time.
3. I use iKnow to Learn items. Once 'mastered' in iKnow, it goes to Anki for long term study and review.
4. The cost of iKnow seems reasonable. What it buys me is Not having to take the time to configure Anki in the same way for "Learning" mode. My time to do that is worth more to me than the cost of iKnow for a year. Everyone has to make that decision for themselves, though.

The stuff that you can do in iKnow and similar sites more quickly and easily than in Anki seems like it's mostly just fluff and contributes to wasteful overlearning, though studies show that such fluff gives students the impression that they're learning more, even when it's less effective for long-term retention (the same poor metacognitive awareness also lends itself to ineffective ‘dropout’ scheduling where they stop studying items too soon for their purposes).

It also often relies on the frequent misconception that Anki is only a passive post-learning repository, rather than a highly dynamic and customizable framework for active knowledge reconstruction on a spaced schedule.

I do look forward to further tinkering with Anki's selective study/card priorities, and in the future, study groups (2.0), to try and accomplish my goals for chaining cards together (e.g. cycles of conversational turn-taking, textual sequences) and customizing review sessions so that certain card types come up in specific orders (for example, iPlus1, iPlus2, etc.). I also haven't been using templates nearly well enough, but am remedying that with subs2srs (see maturity reminder plugin thread). The card chaining is tricky but Karpicke & Roediger's thoughts on spacing have confirmed my own so I'm not too picky about how the individual card schedules for such groups will be effected. It will just be a shift up to a more macro level for those chains.

The other thing is that I've now incorporated a new type of initial day review schedule that more closely mirrors the most effective methods in spaced retrieval studies, with at least 2 successful retrievals on the first day, spaced out by minutes/hours. At the moment it's tricky, but I believe resolve's also going to make this aspect more flexible.

@nest0r Initial day review schedule, what is it exactly?

@jhenson "configure Anki in the same way for "Learning" mode." I've come across this a few times before, like people asking how to learn items for later reviewing in Anki. What I have done in the past is simply review all the items, the ones I know with the ones I don't know and mark the new ones as wrong until I am confident I learned them (at least for the study session).

EDIT: Try the Anki 'Learn Mode' plugin.

Skip to the bottom of this comment to view the simple final version (for now)

Well, I'm still experimenting, but here's the background of my initial thoughts: http://forum.koohii.com/showthread.php?p...#pid141578

As it so happens, a batch of cards I learned with this method just came up in Anki, and I must say, I was a bit elated at the results. In fact, I now regret grading them with 3 instead of 4 (it was out of habit).

At any rate, what I'm going to do with the next batch is slightly different in terms of workflow in Anki (I'm only doing this with vocabulary cards created using Rikaisan during reading, at the moment, still using other strategies for other card types).

1) Study these vocabulary cards using my usual methods of rapid spontaneous mnemonics as needed, and multisensory cues on the front, meaning on the back. In order to cycle through them, just studying, I fail them, then undo those fails. That seems quickest. Edit 4: Correction, I'll do what I did the first time and reschedule them as new, seems easier than hitting ctrl-z 20+ times. ;p

2) A few minutes later, test the cards I just studied. (Edit: It occurs to me I ought to perform a distractor task in order to offset the fact that I'll be testing them in the same order.) Per day scheduling is disabled, with initial button 2 interval (Settings → Deck Properties → Advanced) set to 0.1 to 0.2. This will bring them up a few hours later when passed as Hard this first time. Each card must be passed once, so failed cards are, well, failed, and restudied at the end of the session, then retested; repeat this till all cards are passed once in that session. Edit 3: It occurs to me I missed a step. When failed cards are restudied, I guess I must go to the browser, search for is:failed, and reschedule them as new in order to retest after a few minutes and grade them to push them forward with the other passed new cards. Kind of a pain.

3) A few hours later, test them again. I haven't experimented with this (last time I just rescheduled as new a few hours later), so hopefully when I pass them this second time, I can schedule them for a substantive amount of time later (still messing with this, but I think a max of 24 hours is good. And from there it goes into days, weeks, etc.).

I may also try an extra review that first day if I'm doing many cards. I'm using less than the usual 3 spaced minutes apart because of the extra encoding strategies with richer information per card. Balancing quality and quantity. I got the impression Roediger, et al., had the same feeling, which is why in one paper they recommended two retrievals spaced apart by minutes and hours (with a subsequent delay of two days) rather than three spaced by seconds and minutes. I may try waiting two days as well, instead of 24 hours. Hmm. Actually they recommended two in one example, and one in another. So maybe I'll go with 36 hours. ;p

Edit 2: Also keep in mind feedback, a review of the Answer side of the card, is performed, each time, pass or fail. For passed cards, it's to take advantage of reconsolidation, for failed or difficult cards, it's corrective, including for low confidence/metacognitive errors. (This is also why I've an interest in strategies for dynamic card alterations over time beyond cues, but I digress.)

Edit 5: I wrote up a simpler version without all the thinking aloud, but I want to wait and test a bit before posting it. It also occurred to me that I'm going to be handling all failed cards differently from now on in order to be able to restudy them and then test and grade them. It never mattered before, because I just restudied failed cards, then fake-graded them as Hard so I could test them the next day or so. Now that I figured out a way to divide them properly, it's no longer necessary.

Edit 6: Also, the only reason I've been resetting cards is because I'm uncertain of how Anki handles failed cards; I don't want the initial process of deliberate, ungraded failures to negatively impact their scheduling. The more I read from random forum postings via Google, the more I think it's okay to just fail them.

Edit 7: (Final edit. ;p) Just made another glaringly obvious workflow improvement, so yes, the new version is much more streamlined.

______

Final Version:

1) Study new cards using preferred memorization strategies, via the card browser preview (ctrl-f, search for ‘is:new’, F2 to preview selected card). If you need to use the JDIC plugin, etc., (i.e. you added through Rikaisan), then study as you add audio, hitting button 1 each time, then reschedule them as new cards.

2) Wait a few minutes, then test the cards, making sure to always read the other side and get feedback, even when passing cards.

3) Passed cards are graded Hard to show up in a few hours. (Set initial button 2 to .1-.2, disable per day scheduling.)

4) Failed cards are restudied in the card browser (is:failed, F2), then tested as before; repeat till all from Step 1 have been passed once and thus pushed a few hours away.

5) When they show up for retrieval hours later, test as before, but now grade them so they're due the next day or the day after that. Failure results in Step 4 being repeated, with this new goal of passing and grading for the next day or the day after that. So now you've performed two successful recalls of each card on this first day.

6) That third recall day(s) later and thereafter, proceed as usual, spacing days and weeks apart, etc. I think that method of studying failed cards and new cards via browser first then testing them is good, so I'm going to do that for all cards from now on.

____

Final Final Final Version

1) Study, failing as you go. Reset as new, so leech threshold and review count are unaffected.

2) Test with feedback, passing and failing. What you pass with doesn't matter, since you'll be resetting again in Step 5.

3) Restudy failed in the browser via card preview. Reset as new, so leech threshold and review count are unaffected.

4) Test failed, passing and failing. Repeat 3-4 till all are passed once.

5) Go to the browser, sort by created, select and reschedule this batch of new cards as new again, so you can use the initial button interval in Step 6.

6) Come back in a few or several hours, test with feedback and fail with restudy in the browser as necessary (Steps 3-4), making sure all cards eventually passed with initial button 3, set to .75-1.75. Alternately, you could have graded with button 2 in Step 2, set as .1-.2 for the initial interval; skipped Step 5 and after this step rescheduled for the next day if the new buttons didn't allow for .75-1.75 spacing.

7) Review the next day or so, passing as normal; I think first restudying failed in the browser, then testing them, is best from here on out.

The trouble with ANKI is that you have to learn ANKI before you can learn Japanese. I know some folks will say that Anki is easy to set up any way you like, but it's only easy if are the type of person who like to configure tools as much as use them. Also, I can't get by how ugly it is visually.

That said... it is an undeniably powerful and customizable offering for free.

iknow is way more of "hit the ground running" thing and it looks and sounds are satisfying. I don't find the cost prohibitive, but if others do.. that's a fair critique. When you compare it to the cost of textbooks or a week course from a local language school, I think it's quite competitive.

Sounds interesting. Poor initial memorization may slow down the first steps of reviewing quite a bit. But has anybody tried this practically?
Is it convenient/safe to play this game with one's main anki deck—i.e. changing rescheduling intervals, resetting cards, etc. Does it not interfere with the normal handling of longer term reviews? Would it make sense to maintain a second copy of the deck with all cards suspended except those currently being crammed?

jmignot Wrote:Sounds interesting. Poor initial memorization may slow down the first steps of reviewing quite a bit. But has anybody tried this practically?
Is it convenient/safe to play this game with one's main anki deck—i.e. changing rescheduling intervals, resetting cards, etc. Does it not interfere with the normal handling of longer term reviews? Would it make sense to maintain a second copy of the deck with all cards suspended except those currently being crammed?

Are you referring to the method in my post? Yes, as I've mentioned, it works amazingly. I wrote up a hopefully better explanation on the wiki—see the Website link next to my name here. I'm still polishing the formatting, though, as I learn the wiki stuff. There's no need to create a new deck, I designed it to mesh with and augment normal reviews/cards. In addition, in new versions of Anki it'll be even easier to manipulate small intervals and customize study sessions.

I did not know about this entry in the Wiki. It looks pretty detailed so I should be able to manage. Thanks for taking the time to put together this tutorial !

No problem. I finished polishing it and updating it, I think that's all I can do till Anki 2.0 or whathaveyou becomes official. So from now on when rambling about new cards, instead of linking to the above tangential comment and its rambles and multiple versions, it'll just be: http://rtkwiki.koohii.com/wiki/nest0r%27...ds_in_Anki

Why did I not start using the wiki before? It's so much fun.

wccrawford Wrote:iKnow is the only one I've found that repeats the same word over and over during the study sessions. It'll do it from JP->EN, EN->JP, Kanji->Typing, Audio->EN, Audio->JP, Audio->Typing, Audio Sentence->Typing... I think that's the full list. This repetition is the best way for me to learn a new word.
All the others work basically like Anki. If you get it wrong, it shows up again soon, but if you get it right, it goes away for a while. Sometimes a long while.

This.
This is the crux of the matter. From what some other users have said above it sounds like you could configure anki to do all this, but frankly the iknow interface is really nice and it's already configured to ask you all those questions within the same session, as well as actually teaching you the material by an initial cram.

Definitely worth the money in my opinion.

I haven't used iknow/smart.fm since they went pay, but I found anki far superior to that service due to the frustration dealing with the bugs in iknow I was constantly pointing out and never getting a reply for. If anki was pay and iknow was free, I would still prefer anki.