New 2.0.1 Card Selection Mode

[mikelove]

2.0.1b1 is rapidly approaching release, but there's one thing I'm still weighing the merits of adding which I wanted to ask about here.

It's a new Card Selection option tentatively called "Vocabulary Builder," and basically it's designed to (finally) help with the problem of importing a bunch of new vocabulary and then having too many cards to review at once. (the category "Split" option was supposed to do this, but it's a bit inelegant and nobody seems to be using it much anyway) The way it does this is by limiting the number of "not yet learned" (which probably means they're below a user-defined score threshold) cards that it'll make available in sessions at any given time; if a card has never been studied before, it won't be brought up for the first time until the pool of "not yet learned" cards is below a user-defined size (say 50). There'd likely also be a preference to choose in what order cards would be added to the pool - randomly, by order created, or by a mix of the two (so that cards created on a given day would all be added before cards created on the next day, but not in the exact order they were added, saving you from having to slog through 50 "an"s and "ba"s before getting to other letters of the alphabet in a big list like HSK). There might also be a setting to choose what ratio of new versus old cards you'd like to study in each session, though that's still up in the air (and would probably be too difficult to implement with Endless sessions, so it wouldn't be applied in those, at least not in 2.0.1).

In theory this concept could work for both frequency-adjusted and repetition-spaced sessions, so since the code would be almost the same we'd probably try to offer it in both. It's not actually that difficult to code, either. The big question, however, is whether it makes enough of a difference at all to be worth implementing - the category system already offers the ability to gradually add small groups of new words to your testing pool, so in some respects this is just a lower-maintenance way of achieving the same thing, but that alone might be enough to make a big difference for new users who don't want to take the time to learn the ins and outs of the category system - if we make this the default setting, then really all somebody needs to do is import a list of vocabulary words they want to learn and the software will take care of the rest, no need to organize them at all unless there's a particular reason (like textbook chapters) to do so.

 

[mfcb]

sounds great !

although i am a user of the "split" command (hehe, yes, available features will be used) this sounds even more convenient...

one wish i have to add, please let us choose a category that is treated with high priority in the pooling process. the reason for this is, that i infrequently add parts of HSK to my vocab, but daily i add new words i encounter to my "current" category ("200812"). so what i mean is that i would prefer to add all stuff from "200812" first, before adding anything from HSK...

 

[gato]

one wish i have to add, please let us choose a category that is treated with high priority in the pooling process. the reason for this is, that i infrequently add parts of HSK to my vocab, but daily i add new words i encounter to my "current" category ("200812"). so what i mean is that i would prefer to add all stuff from "200812" first, before adding anything from HSK

This sounds like can be accomplished by making your "current" category the "default category." In Flashcards->Manage Categories, you can make a category the "Default" near the bottom of the screen.

 

[mfcb]

This sounds like can be accomplished by making your "current" category the "default category." In Flashcards->Manage Categories, you can make a category the "Default" near the bottom of the screen.

i know. i add cards to my default category, but my request was to PREFER the new cards from there over the cards from any other list that will be added (or pooled). lets say i add HSK level 3 vocab (approx 2000 cards) to my vocab, then it takes ages until i see vocab i add via the default category over the next few days... its maybe valid to assume, that the preferred category always is the default category (at least for my usage), but maybe not...

 

[renovator]

2.0.1b1 is rapidly approaching release, but there's one thing I'm still weighing the merits of adding which I wanted to ask about here.

It's a new Card Selection option tentatively called "Vocabulary Builder," and basically it's designed to (finally) help with the problem of importing a bunch of new vocabulary and then having too many cards to review at once. (the category "Split" option was supposed to do this, but it's a bit inelegant and nobody seems to be using it much anyway) The way it does this is by limiting the number of "not yet learned" (which probably means they're below a user-defined score threshold) cards that it'll make available in sessions at any given time; if a card has never been studied before, it won't be brought up for the first time until the pool of "not yet learned" cards is below a user-defined size (say 50). There'd likely also be a preference to choose in what order cards would be added to the pool - randomly, by order created, or by a mix of the two (so that cards created on a given day would all be added before cards created on the next day, but not in the exact order they were added, saving you from having to slog through 50 "an"s and "ba"s before getting to other letters of the alphabet in a big list like HSK). There might also be a setting to choose what ratio of new versus old cards you'd like to study in each session, though that's still up in the air (and would probably be too difficult to implement with Endless sessions, so it wouldn't be applied in those, at least not in 2.0.1).

In theory this concept could work for both frequency-adjusted and repetition-spaced sessions, so since the code would be almost the same we'd probably try to offer it in both. It's not actually that difficult to code, either. The big question, however, is whether it makes enough of a difference at all to be worth implementing - the category system already offers the ability to gradually add small groups of new words to your testing pool, so in some respects this is just a lower-maintenance way of achieving the same thing, but that alone might be enough to make a big difference for new users who don't want to take the time to learn the ins and outs of the category system - if we make this the default setting, then really all somebody needs to do is import a list of vocabulary words they want to learn and the software will take care of the rest, no need to organize them at all unless there's a particular reason (like textbook chapters) to do so.

Go for it. Great feature that would make a huge difference!!!! This would be great as I now go through a 3 step process that does not work well in all cases. If I import a list with 3000 words, lets say then I will take the first 25 words and add them to a new category and remove those 25 words from the 3000 list category. Then I learn those words in a Frequency-Adjust mode until I know them well. Once well known, then they are moved to my master list which is Repetition based. The problem comes that some large lists contain words or characters in order of importance or of frequency of use. These lists get all screwed up because in the Manage flash cards section I sort everything alphabetically by pronunciation to be able to quickly find certain cards. It appears that your proposal would both solve this problem as those lists could easily keep their initial sort, and streamline the whole process. Another important feature would be a drop down box to select how many cards to move from the big list. These should be a net cards moved. What I mean by this is that if I select to add 25 cards, it really does add 25 new cards, not say only 20 cards because it finds 5 duplicates in my master list (just skip over any duplicates).

Split works well for small lists of 200 words or less but is too much of a pain to use trying to break down very large lists.

 

[stephanhodges]

Considering the much larger user base you now have (potentially), I would suggest trying to generate fairly good "tutorial level" documentation at the start, even if it means a little extra editing as things go along. Or, we could use my wiki for that if you would want to encourage (beta people) to help on the documentation editing front. This is strictly "tutorial stuff", so should not interfere with your requirement for all user manual material to be generated in house. I can certainly supply all screen shots for Palm, for example.

The suggestion to use the wiki here is to both save time and get more detailed tutorials on this feature. It would take a "little" encouragement, as I've only had 3 others contribute any editing at all so far.

Second, for me it's a little hard to evaluate the proposal without knowing what else is already in the .01 release, since I have already used up 3 levels, and my bottom levels (as mentioned) are averaging about 150 cards apiece, which is too much for me. I admit I'm a complete looser when it comes to how many cards I can learn at a time...

Last, I'd like to encourage you to consider these small number changes as just that, small. Much better to get some items out than to do a "mega" minor update that doesn't actually emerge until next May... :shock:

 

[mikelove]

mfcb - prioritizing categories won't be possible in the initial release, unfortunately, but we can certainly look at adding it as an option in the future if this feature proves popular - for right now the only options would be choosing cards randomly and by creation date.

renovator - good point about splits being impractical for large categories. This doesn't involve any work on the user's end, though, you don't have to manually move files between lists - the software just checks at the start of every session to see how big the pool of un-learned cards is, and if it's below the desired level it adds a few new (previously un-reviewed) cards to the pool to bring it up to that level.

stephanhodges - actually what I'd really like to do on that is a set of tutorial videos; works much better than a static page, and we can probably even attract some new customers by posting them on YouTube. And I agree with you completely on keeping the .0.x releases small, we can't fix all of the rough areas in 2.0 in one patch, but at the same time after 2.0.1 it could be quite a few months before we update the Palm/WM versions again (since we'll really be going full-bore on iPhone) so I'd like to fix as many little things as possible.

 

[cming]

There might also be a setting to choose what ratio of new versus old cards you'd like to study in each session, though that's still up in the air (and would probably be too difficult to implement with Endless sessions, so it wouldn't be applied in those, at least not in 2.0.1).

Great new feature! Would be excellent if you could choose the ratio of new and old cards to be reviewed, I didn't know the split option could acheive a similiar effect, I'll certainly give that a try, but since most of my lists are chapter based, it might be of limited use. Giving priority to a category on a permanent basis would be useful for me, for example I have a "Names" category, which I add the names of everyone Chinese I meet, and I really have to know these ASAP. At the moment I just turn off all other categories when I want to review the "Names" category, but it would be great to be able to mix them with the rest of the cards, with new cards in that category, getting priority over new cards added to another category. Even if this comes in a later version of Pleco, it will certainly be worth waiting for!

 

[cming]

Conversely, if this feature is successful, you could also look at making the whole flashcard system a bit more intuitive for learning new cards - I think this would especially benefit Elementary/Intermediate learners. Starting with the simplified character, then moving to pinyin/zhuyin, then stroke order, example sentences, and finally the traditional form of the character (if there is one). [I'm sure someone will point out here that there is a good pedagogical reason for learning the traditional form of the character first]. I know you can do all of these things seperately now, but moving through them in a more step-like fashion could be useful. In other words you would focus on learning the whole card step by step, before moving onto the next card.

Testing could also be slightly expanded: for example, in a free answer session showing "audio" why can you only ask either "Headword" OR "Pronunciation", and not both?

 

[daniel123]

The big question, however, is whether it makes enough of a difference at all to be worth implementing

Yes, for me it makes a big difference. Would be great if you can implement it.

Daniel

 

[stephanhodges]

Any idea where along this timeline we might get testing for individual characters as well? (Not generating additional cards in this case). Also, would be nice to be able to find cards that USE the character/pronunciation combination.

 

[mikelove]

cming - it looks like that ratio setting might make it in, but only when using Random card selection - Frequency-adjusted already does the same sort of thing in a more fine-grained way, so there's not much point in adding it for that (and it would be a lot of work), and with Repetition-spaced there's a fixed set of cards that you need to study on a given day so it really doesn't make sense to delay / drop out some of them in order to preserve that ratio. (the main limit-unlearned option will be available in both modes, though, you just won't be able to set the ratio)

And there's certainly a lot of room to allow for easier and more flexible selection of which parts of a card you're tested on - we've doubled the number of score-based field selections in 2.0.1 (you get up to 8 now instead of the earlier 4) but it's still not very intuitive and it still doesn't let you mix, say, free-answer and multiple-choice in the same test.

stephanhodges - I don't think I really like the idea of testing individual characters in cards, actually; a lot of aspects of the flashcard system are based around the idea that cards are atomic / can't be broken down further, and violating that both makes things more confusing and creates a whole lot of new design problems (how to adjust scores if you get one character correct and the other incorrect, how to avoid having the same single character come up too often because it appears in a lot of words, etc). Plus, modern Chinese has a whole lot of bound characters that there's almost no reason at all to study individually.

You can search for cards with a particular character / pronunciation combo in Manage Flashcards if you just AND a "headword" with a "pron" search - is that what you're referring to?