#163 Timestamp Self Review Incorrectly updated

[Michael van den Berg]

Using the 'Interval Filter' when studying vocab (in tab Flashcards -> Click green arrow, then under options), incorrect answered flashcards are not always presented the next time the same set (with the same setting) is studied.

What I mean is this:

• There is a set of Flash cards (about 10-20 categories)
• I study, for example 5 categories (any number will do), in total about 30 flashcards (also here any number will do) with 'interval filter' turned on
• Go through the whole deck at least once - a number of cards are answered wrong, some correctly. Say, there are 12 cards left over.
• If desired, keep studying until only a few cards remain, but doesn't seem to make a difference for this bug.
• Finish the sessions by clicking the Stop button or Start button in the top-right corner.
• Start a new study session 1 minute after the first, with exactly the same settings. At least 12 cards should be in there, because 12 cards were NOT answered correctly and need review. Surprisingly, less than 12 cards, for example only 8, remain.

I believe this is what is happening:
1 It was long ago that these cards where studied. The 'score' was quite high, e.g. 5 (this implies that the cards was answered correctly at least 5 times, more or less, in a row)
2 When reviewed, the time stamp on the flash card is reset to the current time, let's say 15:01.
3 When the card is correctly answered the 'score' is incremented with 1 (which is correct).
4 When the card is answered incorrectly, the score is updated minus 1 (which is also correct).
However, now the cards that were wrong, are at score 4 (if the previous score was 5). The time stamp is set to now (15:01). At 15:02 same day, we start a new session. The interval filter tells that at this score of '4', ZDT will wait 4 days before presenting the card. This shouldn't happen this way. If understood correctly, the time stamp should only be updated when the card is answered the first time correctly.

I am not 100% about the exact algorithm, but it seems to me something like that.
For example, I don't know how do we differentiate between a situation were the first time a card is answered incorrectly, but by the second round (during the same session) is answered correctly? Or do we get exactly the same statistics for a card that is answered 10 times wrong as for a card that is answered only once wrong?

What I do know is, that if a card is not answered correctly, I should definitely get it again in any next session where I select that category. And that doesn't happen now.

I've attached a ZDT backup, with a number flashcard entries that made me discover this behavior.

This touches the core functionality of ZDT. There is not really a workaround, except turning off the 'interval' filter, which partially defeats the purpose of using software.

[Michael van den Berg]

Oliver, do you think you could find how to fix this bug?
If you can and you have a .JAR file ready, I am happy to test that one.