Topic closed
Reply #851 - 2007 October 28, 10:02 am
Aikiboy Member
From: Kitakyushu Registered: 2006-12-09 Posts: 127 Website

Hi Resolve, my cell still shows the same error, after the first page and enter, it freezes.  Midnite, Tokyo time.

Reply #852 - 2007 October 28, 10:36 am
resolve Member
From: 山口 Registered: 2007-05-29 Posts: 919 Website

Some cellphones cache the flash file. Did you force a refresh?

Reply #853 - 2007 October 29, 9:04 pm
Killersquierl New member
From: Wisconsin Registered: 2007-05-04 Posts: 8

Hey resolve, I have an idea for a new feature. It would be cool if cards could be dependent on other cards: that is to say, until the required cards are 'learned', that card wouldn't be introduced. Then you could just make a giant deck that introduces new cards when the time is right.
  Take for example RTK1 and RTK2. You could set it up so that each kanji from RTK1 is a card and that each group from RTK2 is a card. The thing is that each group from RTK2 wouldn't be introduced, until the proper kanji have been learned from RTK2. So you could have the cards in the deck, but Anki wouldn't show them until you've met the requirements.
  Taking it a step further, you could make each complex kanji in RTK1 have as requirements, the primitives that kanji contains. For example: anki wouldn't let you learn the kanji for risk until you had a learned the ones for sun and ear.
  Of course one question that comes up is this: When is a requirement fulfilled? Is it after you recalled it correctly x number of times, or would it based of the repetition time period? That would require some thought.

  Hypothetically speaking, one could have a deck with basically three cards. More if you're not just doing RTK.
   *First would be the primitives/kanji that are primitives.
   *Second would be kanji that depend on primitives.
   *Third would be RTK2 reading groups dependent on kanji.
  Other ones that aren't RTK, but might want them as dependencies:
   *Fourth would be compounds or words that are dependent on the corresponding RTK2 cards.
   *Fifth would be sentences (for the AJATT method) that depend the correct words       
    being known, along with the correct particles and grammatical concepts.
   *Sixth would be particles, pure katakana or hiragana words; grammar concepts

  Of course the problem with this is that after a while, a lot of cards would be waiting to be seen. You might learn a concept early and not get to the things that build off it forever. Also, you might want to learn a specific card, but it's not allowed. The last thing needed to round out this feature would be the ability to prioritize cards to be allowed. So someone could say I want to learn this card; and then Anki would show the requirements for that card first, so the card the person wants to learn is introduced as fast a possible.

To sum up there would categories cards would fall into if this were a feature:
-Scheduled for review
-Waiting to be reviewed
-Waiting to be seen for the first time
-Waiting to be allowed; these cards won't have a chance to be seen until their requirement are fulfilled.

  That's what I think. It could probably be extended to math or other things to. Personally, if it is going to implemented, I think it should be a special feature for the field. So you could add a field and then under 'Special features used by this field' you would just put the word requirements or something. Then in the field you could list the requirements. Someone could also (I don't know if this is possible) make an requirement-auto-analyzer as a plugin.

  I just think it would be really cool and useful.

Advertising (register and sign in to hide this)
JapanesePod101 Sponsor
 
Reply #854 - 2007 October 29, 10:03 pm
resolve Member
From: 山口 Registered: 2007-05-29 Posts: 919 Website

Thanks for taking the time to write such a detailed request.

Unfortunately, I'm not sure this is such a great idea. It would:
a) complicate the interface
b) require a lot more work to actually add material

If someone were making Anki decks for money, they might have the motivation to go to the trouble of listing the dependencies. But dependencies are just a round-about way to ensure cards are displayed in a particular order. You can accomplish exactly the same thing by adding cards in a particular order, and making them be randomized after they've been seen for the first time. This is exactly what the Heisig deck does, and it does it much more easily than trying to enumerate all the dependencies of a given card.

Last edited by resolve (2007 October 29, 11:20 pm)

Reply #855 - 2007 October 30, 3:12 am
zazen666 Member
From: japan Registered: 2007-08-09 Posts: 667

Hi Resolve,
Question about using cell phone with ANKI
Do you know how can I check on my docomo phone if anki website is being `cached` with the initial startup ?

Reply #856 - 2007 October 30, 4:21 am
resolve Member
From: 山口 Registered: 2007-05-29 Posts: 919 Website

If you're not sure, open the menu and save the flash file to your phone's data folder. You should then be able to run it from the data folder and ensure it's not being redownloaded.

(It's only 5k though)

Reply #857 - 2007 October 30, 4:29 am
wrightak Member
From: Tokyo Registered: 2006-04-07 Posts: 873 Website

Killersquierl wrote:

Hey resolve, I have an idea for a new feature. It would be cool if cards could be dependent on other cards

You may be interested to read the discussion that I had with resolve here:

http://repose.ath.cx/tracker/anki/issue144

I also thought that card dependencies would be a good idea but after using Anki for a good while now and thinking about this, I've come to the conclusion that the benefits are uncertain whereas the costs to implement are quite great. It's an interesting topic though.

Reply #858 - 2007 October 30, 10:01 am
Aikiboy Member
From: Kitakyushu Registered: 2006-12-09 Posts: 127 Website

Righto on the force refresh issue.  Thanks again! BTW, this week I've taken to running Anki from my usb flash memory.  So much lighter to carry around. smile

Reply #859 - 2007 November 01, 4:26 am
wrightak Member
From: Tokyo Registered: 2006-04-07 Posts: 873 Website

It seems that the online interface now takes into account prioritised tags. I thought that it didn't before or am I just imagining that? Have there been any changes?

Does the online interface handle suspended cards in the same way as the desktop one?

Reply #860 - 2007 November 01, 5:05 am
vosmiura Member
From: SF Bay Area Registered: 2006-08-24 Posts: 1085

When does Anki display card statistics in the "Cards over a week" range?  I've been reviewing with Anki for a couple of weeks and have done many reviews after more than one week but it still shows 0/0 cards answered over a week.

These are the stats form my Heisig deck after 20 days.

Cards over a week: 728 (35.65%)
Cards under a week: 357 (17.48%)
Unseen cards: 957 (46.87%)

Correct over a week: 0.0% (0 of 0)
Correct under a week: 89.8% (2291 of 2552)
Correct unseen: 81.4% (420 of 516)

Average next time: 16 days
Average workload: 68.8 cards/day
Cards added: 22 a day, 668 a month

Period Cards Deck%
1 day 39 3.59%
3-5 days 96 8.85%
6-9 days 160 14.75%
9-21 days 445 41.01%
21-45 days 345 31.80%

Last edited by vosmiura (2007 November 01, 5:14 am)

Reply #861 - 2007 November 01, 5:44 am
resolve Member
From: 山口 Registered: 2007-05-29 Posts: 919 Website

wrightak: The web version is using the development version of the Anki library, which chooses the next card to display in a different order. Both the old web interface and new web interface honoured suspended and prioritized cards.

vosimura: Mature cards are cards that have been successfully remembered for a level 4 interval. So when you answer a card that has an interval of 9 days, before you answer, it is young, and after you answer, it is mature. Thus, for the purpose of statistics, the correct/incorrect answer is treated as an entry in the young statistics.

So mature cards are cards that have an interval > 9, using the default settings.

As for why that number is zero, that is indeed odd. Could you send me your deck?

Reply #862 - 2007 November 01, 5:58 am
wrightak Member
From: Tokyo Registered: 2006-04-07 Posts: 873 Website

So the web version is using the algorithm that you introduced and we discussed earlier in this thread?

Reply #863 - 2007 November 01, 5:58 am
resolve Member
From: 山口 Registered: 2007-05-29 Posts: 919 Website

yes

Reply #864 - 2007 November 01, 8:12 am
johnzep Member
From: moriya, ibaraki Registered: 2006-05-14 Posts: 373

resolve wrote:

As for why that number is zero, that is indeed odd. Could you send me your deck?

I've noticed that the "Correct over a week" seems to be low too.

Deck created: 92 days ago
Total number of cards: 2586

Cards over a week: 1710 (66.13%)
Cards under a week: 561 (21.69%)
Unseen cards: 315 (12.18%)

Correct over a week: 97.6% (205 of 210)
Correct under a week: 69.7% (1600 of 2294)
Correct unseen: 45.9% (184 of 401)

Average next time: 39 days
Average workload: 58.2 cards/day
Cards added: 28 a day, 845 a month

Period
Cards
Deck%
1 day 149 6.56%
3-5 days 53 2.33%
6-9 days 182 8.01%
9-21 days 455 20.04%
21-45 days 527 23.21%
1.5-3 months 809 35.62%
3-6 months 94 4.14%

There are 900 cards with intervals over 1.5 months, but it's only counting 205/210

Reply #865 - 2007 November 01, 1:48 pm
vosmiura Member
From: SF Bay Area Registered: 2006-08-24 Posts: 1085

resolve wrote:

As for why that number is zero, that is indeed odd. Could you send me your deck?

Heisig deck for "vmiura" on the server.  Or do you need my local deck?

Reply #866 - 2007 November 01, 2:35 pm
resolve Member
From: 山口 Registered: 2007-05-29 Posts: 919 Website

I looked at your deck. Everything seems to be fine.

Perhaps "cards over a week" could be rephrased to be clearer. At the moment you have no cards over about 23 days. You scheduled a card at 'ease 4' and it gets an interval of, say, 8.4 days. The next time you answer it, it's given an interval of, say, 23 days. But the card is a 'young' card at 8.4 days, so the stats go in the "under a week" section. Once you answer the card again in 23 days, they will go in the "over a week" (mature) section.

Reply #867 - 2007 November 01, 8:41 pm
vosmiura Member
From: SF Bay Area Registered: 2006-08-24 Posts: 1085

resolve wrote:

I looked at your deck. Everything seems to be fine.

Perhaps "cards over a week" could be rephrased to be clearer. At the moment you have no cards over about 23 days. You scheduled a card at 'ease 4' and it gets an interval of, say, 8.4 days. The next time you answer it, it's given an interval of, say, 23 days. But the card is a 'young' card at 8.4 days, so the stats go in the "under a week" section. Once you answer the card again in 23 days, they will go in the "over a week" (mature) section.

Hmm, so I have to answer '4' 3 times in a row?  Anyway, just checking.

johnzep's statistics look more odd though; he has a lot of cards over 1.5 months but a lot less in the over a week section.  I guess he would have had to answer '4' at least 3 times on those cards already.

Edit: Come to think of it, all his review counts seem low compared to the number of cards at high intervals he has.  Does it count cards reviewed on different hosts?  johnzep, did you add a lot of cards and change the start interval when adding them?

Last edited by vosmiura (2007 November 01, 8:51 pm)

Reply #868 - 2007 November 01, 9:41 pm
johnzep Member
From: moriya, ibaraki Registered: 2006-05-14 Posts: 373

I've never changed the intervals.  The deck is made up of the 3級 and 2級 sample decks included with anki in addition to words and sentences I've added.

Maybe one source of the oddities is from merging the decks...before their were the 2 sample decks, my own vocab, and my sentences all in separate decks.  I merged them all by using the online syncing.

resolve Member
From: 山口 Registered: 2007-05-29 Posts: 919 Website

Before 0.3.0, statistics weren't synced, so each computer kept separate statistics. That's probably it.

Reply #870 - 2007 November 04, 1:31 pm
uberstuber Member
Registered: 2007-03-27 Posts: 238
Reply #871 - 2007 November 04, 2:47 pm
resolve Member
From: 山口 Registered: 2007-05-29 Posts: 919 Website

Not really :-) When I choose the name 'Anki', I googled it to see if it was already being used, and found the palm software, but the project was dead. Other names I'd been considering but decided against were Choanki (think 超暗記)  and AnkiNg (pronouced AnKing)

Reply #872 - 2007 November 04, 4:23 pm
HerrPetersen Member
From: Germany Registered: 2007-01-02 Posts: 238

I actually downloaded it and used it a little, always thinking, it was yours.

Last edited by HerrPetersen (2007 November 04, 4:24 pm)

Reply #873 - 2007 November 05, 9:01 pm
dilandau23 Member
From: Japan Registered: 2006-09-13 Posts: 330

I have been thinking about the lack of a proper Anki forum.  It is really getting troublesome to read through 30+ pages of posts just to check if the question I want to ask has already been answered.  I know there is the Anki bug tracker and the faq.  I was thinking, since this is someone else's site anyway, could we at least get a "question" priority in the bug tracker if nothing else?

Reply #874 - 2007 November 06, 3:35 am
Weber Member
From: Hanover, NH Registered: 2006-12-30 Posts: 22

Hello everyone,

I just started using Anki today, and I can't wait to get started, it seems to be a great complement to this site for reviewing grammar (particularly for JLPT 1級) and vocabulary. Unfortunately, I'm at university now, and much to my chagrin I just can't afford to spend the large amount of time necessary to make the kind of decks I want for my studies. I just finished reading through this whole thread, and found some JLPT .anki decks that Resolve had made available, but they no longer work with the new versions of Anki (Im running 0.3.6.) I would really appreciate it if some people here could post/email any decks they have for JLPT 1 and 2 (vocab, grammar, etc.), and any decks of sentences they have made. Could you share them with me, please? Thanks VERY much.

Also, I noticed that JimmySeal had made a deck (it might be for another program) of words from novels, and possibly sentences too. Jimmy, if you could share those with me I would be ecstatic, and much obliged. I apologize that I do not have anything of my own to contribute at the moment. I hope to hear from some of you soon.

Reply #875 - 2007 November 06, 3:57 am
vosmiura Member
From: SF Bay Area Registered: 2006-08-24 Posts: 1085

Hi Webber. Have you tried the sample decks that Anki installs? There are some JLPT decks to get you started.

By the way, Anki's "suspend fact" feature may be useful to you in skipping Heisig kanji that you don't want to review. I use Anki for RTK.

Last edited by vosmiura (2007 November 06, 4:02 am)

Topic closed