Pleco for Android 2.3.11-14 Bug Report Thread

[Wan]

Re: Flashcard session crashes

I had something like this recently as well. But it didn't crash per se, just kind of reset and regenerated a new set of cards to review. It's only happened once though and it was a fairly large session (over 500 cards to review).

Yep, that’s what I’m talking about (I think). It hasn’t occured since, but what actually happens from time to time is that when a flashcard session is saved and I then try to continue it, from tapping „Resume Test In Progress“, I get to the screen saying „Begin Test Session”.

I installed alogcat and will send you the logfile as soon as this bug occurs.

 

[mfcb]

glitch in search interface: when searching for a date the selected date shows up just after change of >/<, but it seams that the search uses the right date anyway...

 

[mfcb]

when i open pleco with the flashcard shortcut i did not expect it to open the reader with the current clipboard content...
also, i remember you said that when back out it would not back to the dictionary, in the "next" version, hehe

 

[mikelove]

Yep, that’s what I’m talking about (I think). It hasn’t occured since, but what actually happens from time to time is that when a flashcard session is saved and I then try to continue it, from tapping „Resume Test In Progress“, I get to the screen saying „Begin Test Session”.

I installed alogcat and will send you the logfile as soon as this bug occurs.

OK, thank you!

glitch in search interface: when searching for a date the selected date shows up just after change of >/<, but it seams that the search uses the right date anyway...

Are you sure you're using the latest version? Though we'd fixed this one several releases ago.

when i open pleco with the flashcard shortcut i did not expect it to open the reader with the current clipboard content...
also, i remember you said that when back out it would not back to the dictionary, in the "next" version, hehe

Yeah, we ended up putting a bunch of other stuff in that update so it got backed up a bit... but actually now thanks to Android 4.0 most people switch apps using the fast app switcher instead of by going back to the home screen, in which case you'll be taken right back to where you were regardless - are you running an older version of Android or do you prefer to switch with the home screen even on 4.0?

 

[mfcb]

using pleco 2.3.14 on android 4.0.3 galaxy s2

updated to 2.3.14 on the day it was available in google play...

remark: as much as i was a fan of hardware buttons/keyboard on WinMo, now i am not. i avoid pressing the home button (and any other) as long as i can, as i have the feeling its still much faster to press back several times than to wait for the taskmanager to come up (on long press menu button)...

 

[mikelove]

using pleco 2.3.14 on android 4.0.3 galaxy s2

updated to 2.3.14 on the day it was available in google play...

OK, we'll see if we can reproduce this / fix it again, then.

remark: as much as i was a fan of hardware buttons/keyboard on WinMo, now i am not. i avoid pressing the home button (and any other) as long as i can, as i have the feeling its still much faster to press back several times than to wait for the taskmanager to come up (on long press menu button)...

That's the issue, actually, you're using Android 4.0 on a phone designed for Android 2.3 - on Android-4.0-specific devices there's a dedicated onscreen task manager button that comes up with a regular tap. But plenty of other people are doing that too, so it's certainly worth looking into other ways to handle this.

 

[ziggy]

Hello,
When testing myself with flashcards, I sometimes receive a "red" (wrong) response to my pinyin writing whilst it seems right to me.
Two examples here below : I did make a mistake for the first character, but for the last one, the pinyin is right - or ?..

 

[mfcb]

edit： -----

 

[mikelove]

Hello,
When testing myself with flashcards, I sometimes receive a "red" (wrong) response to my pinyin writing whilst it seems right to me.
Two examples here below : I did make a mistake for the first character, but for the last one, the pinyin is right - or ?..

This is a side-effect of the fact that we actually let you type more than one syllable in a single field (people occasionally use weird input methods or copy-and-paste or for whatever other reason end up entering multiple syllables) - our Pinyin parser is interpreting the "kwai" as "k + wai" and the "jao" as "j + ao" and hence considering those to be the first two syllables of your answer, so that you end up with both of them marked incorrect. (single letters are considered valid Pinyin syllables because of the large number of contemporary expressions that use them - 卡拉OK, B股, BP机, etc)

 

[ziggy]

I see.
If understand you well, if you prevented multiple-syllable word to be entered in a single field, this "glitch" would disappear. I think it would be a good idea, since, when used "normally" (which means "my way" of course), one syllable at a time, the system jumps nicely from one syllable field to the next, as you expect.
Actually, I can't see why you allow entering more than one syllable in a field, since this would inevitably return a "wrong" (red) response - or not.. ?

I have two more questions :
1 - Is it possible to avoid the same word to be in several card categories ? Say, a command doing a screening on all your cards and asking for each "double card" whether you want it in one or another category, or in several categories at the same time ?

2 - Would it be possible to create a button on the main screen that would allow to jump to "resume test", instead of cliking three or four times to get to it ?

 

[mikelove]

If understand you well, if you prevented multiple-syllable word to be entered in a single field, this "glitch" would disappear. I think it would be a good idea, since, when used "normally" (which means "my way" of course), one syllable at a time, the system jumps nicely from one syllable field to the next, as you expect.

Well the question is whether more people are likely to accidentally enter more than one syllable in a field (e.g. due to a wacky Android IME) or to misspell their Pinyin - I'm not sure how we'd go about measuring that but I suspect the numbers are relatively close.

Actually, I can't see why you allow entering more than one syllable in a field, since this would inevitably return a "wrong" (red) response - or not.. ?

No, the idea is that someone might (and indeed some people do) end up with a multi-syllable answer confined to just one field. So you'd have "hao3yi4" in the first field and the second field blank, in which case we'd still mark your answer correct since it would be.

1 - Is it possible to avoid the same word to be in several card categories ? Say, a command doing a screening on all your cards and asking for each "double card" whether you want it in one or another category, or in several categories at the same time ?

It's possible but there's no way to do it after the fact; you can, however, configure the software to always display a duplicate prompt (even when the card isn't in your currently selected category) so that you can decide in each case whether or not to create a new card. Configure that in Settings / Flashcards.

2 - Would it be possible to create a button on the main screen that would allow to jump to "resume test", instead of cliking three or four times to get to it ?

We've added one to the menu on Android 4.0 (where the menu is now a long list and can hence have more than 6 options without burying them in a separate screen) for our next update, though on that you're also taken back to the screen you were previously in when you switch apps.

 

[ziggy]

Excellent, thank you for your response.

I have another example of the software returning a "wrong" (red) response to a correct syllable. It seems to be a different case from the previous one. The first field is empty, whilst the second answer is correct but scored as wrong.

Also, I am not sure that I support your logic to accept a misuse of the separate syllable fields (by putting two syllable instead of one in a given field), whereas the system "penalises" correct answers, for users that correctly follow the rules. No big issue, of course ; it just makes the results a bit worse, which is challenging...

By the way, the scoring system/algorythm is absolutely obscure for my limited intelligence - it seems very smart, but complex ; I therefore try to use some basic parameters and trust the software for presenting the cards at the right frequency and drive my progress.

 

[mikelove]

I have another example of the software returning a "wrong" (red) response to a correct syllable. It seems to be a different case from the previous one. The first field is empty, whilst the second answer is correct but scored as wrong.

Did you enter anything in the first field? If not then it may be another exercise of that same decision to combine the output from all the fields.

Also, I am not sure that I support your logic to accept a misuse of the separate syllable fields (by putting two syllable instead of one in a given field), whereas the system "penalises" correct answers, for users that correctly follow the rules. No big issue, of course ; it just makes the results a bit worse, which is challenging...

It's purely a question of how it marks it in the result fields - for scoring purposes, the system treats an answer with one syllable incorrect exactly the same as one with all syllables incorrect. And it's only a problem when you misspell the Pinyin, or when you insert a single space (or something else space-like) into a field so that it doesn't interpret you as having skipped over it.

By the way, the scoring system/algorythm is absolutely obscure for my limited intelligence - it seems very smart, but complex ; I therefore try to use some basic parameters and trust the software for presenting the cards at the right frequency and drive my progress.

It actually used to be a lot simpler, but then SuperMemo and Anki got popular and people started demanding that we support all of these additional SRS options for it, which meant it could no longer be transparent and straightforward.

 

[ziggy]

Thank you for your answer, but I don't understand. I'm lost now.

You wrote that :
"for scoring purposes, the system treats an answer with one syllable incorrect exactly the same as one with all syllables incorrect"

This is not true in my experience : when I mistype one syllable, and type correctly the other(s), I generally have (as expected) a green field for the right syllable(s) and a red field for the wrong one.

Of course you cannot do statistics on all your users, but I bet that most would consider that a right pinyin syllable in a syllable-field should be scored "right". What else ?

 

[mikelove]

You wrote that :
"for scoring purposes, the system treats an answer with one syllable incorrect exactly the same as one with all syllables incorrect"

Sorry, by "scoring purposes" I meant "deciding when to show you a card again in SRS tests"; i.e., whether a card will be marked in the history as "incorrect" and reviewed again sooner. That part works exactly the same regardless of whether a card is partially or entirely incorrect. So the green versus red is purely a UI issue, it doesn't factor into the underlying algorithm.

 

[mfcb]

not sure if its default settings or something i messed up, but today i reciginezed my scoring sent a card from 10000 to 5000 on answer 3, but on answer 1 (in my opinion the worst case) just sent the card to 7000...

setting called like "incorrect score decrease" but its the destination percentage, so the default of 70% for quality 1 results in score 7000 after incorrect 1, whereas incorrect 3 results in 5000, which really surprised me....

 

[mikelove]

not sure if its default settings or something i messed up, but today i reciginezed my scoring sent a card from 10000 to 5000 on answer 3, but on answer 1 (in my opinion the worst case) just sent the card to 7000...

setting called like "incorrect score decrease" but its the destination percentage, so the default of 70% for quality 1 results in score 7000 after incorrect 1, whereas incorrect 3 results in 5000, which really surprised me....

The defaults for all three are 0, actually, reflecting the standard SRS behavior of dropping a card to the minimum interval as soon as you answer it incorrectly. As for making it a destination versus a subtraction percentage - there are certainly arguments for doing it both ways, we actually picked what you might refer to as the "Chinese way" since in China discounts (on merchandise, say) are described in terms of the destination percentage (打7折 = you pay 70% of the full price) rather than in terms of how much we're subtracting.

 

[ziggy]

by "scoring purposes" I meant "deciding when to show you a card again in SRS tests"... So the green versus red is purely a UI issue, it doesn't factor into the underlying algorithm.

I understand better, thank you.
Still, I consider that having a right pinyin syllable returned as "wrong" (red) is unfomfortable at least, and confusing :? . But I sure can live with it !

 

[Sgolym]

Hello!

I'm an issue with sorting unlearned cards. I'm using automatic SRS, "Limit # of unlearned" is checked, Prioritize by -> "oldest".
The option under "Prioritize by" seems to have no affect, the order is random every time. I've tried taking the same group of cards and resetting all card statistics, as well as trying prioritize by newest, just to make sure the order was changing every time.

I'm not sure if there's some other option that could be taking precedence, I couldn't find it on my own if there is. Any help would be much appreciated!

 

[mikelove]

I'm an issue with sorting unlearned cards. I'm using automatic SRS, "Limit # of unlearned" is checked, Prioritize by -> "oldest".
The option under "Prioritize by" seems to have no affect, the order is random every time. I've tried taking the same group of cards and resetting all card statistics, as well as trying prioritize by newest, just to make sure the order was changing every time.

"prioritize by" affects the order in which cards are introduced, not the order in which they display during the test. "prioritize by -> oldest" means that the cards that you created first will be added to the pool of actively-studied cards first, but they'll show up randomly during actual tests. There's no way to force a particular sort order on cards that you're testing with SRS during the test; it wouldn't really make sense to offer one given how SRS works.

Sorry for the confusion.