2.0.8 Beta 2 Bug Report / Feedback Thread
- Thread starter mikelove
- Start date
still problems with "undo incorrect": it now restores the score, but the card still comes up in "repeat incorrect" and the final session statistics said 2 correct/ 1 incorrect, which was the card i applied undo on. the card count during "repeat incorrect" seems ok now (apart from the undo problem).
HW60
状元
The results of Flashcards/Statistics and the Card Filter Not correct in row have similar results, but they should be identical. My test base is
a) 2:0 ++ (not learned)
b) 6:2 ++-+++-+ (not learned)
c) 5:2 +-+-+++ (learned)
d) 4:2 --++++ (learned)
e, f) 0:0 (not learned)
Flashcard Statistics shows Total 6, Learned 2 which is correct. With not correct in row 3 I get 5 cards instead of 4. Creating another category and putting card c) into that category I found out, that with that card the statistics and the Card Filter show different results. Unfortunately there is no card history (as it was in Pleco 1.0 as far as I remember), so I cannot find out why c) is part of Not correct in row 3.
a) 2:0 ++ (not learned)
b) 6:2 ++-+++-+ (not learned)
c) 5:2 +-+-+++ (learned)
d) 4:2 --++++ (learned)
e, f) 0:0 (not learned)
Flashcard Statistics shows Total 6, Learned 2 which is correct. With not correct in row 3 I get 5 cards instead of 4. Creating another category and putting card c) into that category I found out, that with that card the statistics and the Card Filter show different results. Unfortunately there is no card history (as it was in Pleco 1.0 as far as I remember), so I cannot find out why c) is part of Not correct in row 3.
HW60
状元
I changed cards
e) +---- 1:4 (not learned)
f) +----- 1:5 (not learned),
and now the 4 Card Filters (not) (in) correct in row did what they are expected to do! Maybe I made a mistake in my previous test.
If the information is available and not too difficult to show, I suggest to add a line in Card Info / Statistics like
Record: +-+-++++ (6:2) or
Last 10: +-+-++++ (75%) or just
Last records: +-+-++++,
but as long as Statistics and the Card Filter show the same results, that would only be nice to have.
e) +---- 1:4 (not learned)
f) +----- 1:5 (not learned),
and now the 4 Card Filters (not) (in) correct in row did what they are expected to do! Maybe I made a mistake in my previous test.
If the information is available and not too difficult to show, I suggest to add a line in Card Info / Statistics like
Record: +-+-++++ (6:2) or
Last 10: +-+-++++ (75%) or just
Last records: +-+-++++,
but as long as Statistics and the Card Filter show the same results, that would only be nice to have.
mfcb - I can see the problem with the incorrect count, but I'm not able to reproduce your issue with incorrect cards still coming up in "repeat incorrect" after being undone - are you sure this isn't just a card that came up more than once in a session? And that you didn't wait to undo it until after you'd already started repeating incorrect cards?
HW60 - good point about the history display, you actually can see it if you export your cards in XML format but it'd be nice to squeeze it into the UI somewhere - kind of tricky at the moment, though, the Statistics tab in Char Info is already filled up when you include all of the new buttons.
With your initial Card Filter issue, I tried plugging in those same histories here and the values with Statistics and Card Filter matched, so I'm inclined to think this was just a configuration issue - we definitely need to do more testing on this on our end, though.
Nothing about the compounds user dict code changed in Beta 2 - is there any chance you might have just been looking up a character that didn't have a user dictionary entry?
HW60 - good point about the history display, you actually can see it if you export your cards in XML format but it'd be nice to squeeze it into the UI somewhere - kind of tricky at the moment, though, the Statistics tab in Char Info is already filled up when you include all of the new buttons.
With your initial Card Filter issue, I tried plugging in those same histories here and the values with Statistics and Card Filter matched, so I'm inclined to think this was just a configuration issue - we definitely need to do more testing on this on our end, though.
Nothing about the compounds user dict code changed in Beta 2 - is there any chance you might have just been looking up a character that didn't have a user dictionary entry?
i'm pretty sure, because i had just a feeling at first, thats why i intentionally made a session with just 3 cards, where i answered just one as incorrect. it came up in repeat incorrect... but right now i'm also not able to reproduce it (at least on cards that are not due today, i'll try again tomorrow)...
my repeat settings are loop,shuffle and never save session, if that makes any difference.
my repeat settings are loop,shuffle and never save session, if that makes any difference.
HW60
状元
I do not like the XML export files very much, because they are very complicated to read in Excel or Access (at least for me). So I wait - on a Touch HD the Card Info / Statistics fill only half of the screen ...
My initial card filter problem may have come from using number of cards 5 instead of all - I do not remember. Actually there is no problem with this filter.
The only thing I changed with the user dict in beta 2 is: I created a new user dict in beta 2 and imported cards. I thought that beta 2 creates upgraded user dicts, because I did not find the upgrade button for this dict in manage dictionaries, even after I quit the program, but after a soft reset the upgrade button was there again, and after an upgrade of the user dict this dictionary was visible in compounds again. Therefore I think the user dict can only be seen in Compounds, if the user dict is upgraded.
Two more remarks: is it possible to change the font size in Card Info / Text/Categories? Can you use the full screen in Flashcards / Statistics?
My initial card filter problem may have come from using number of cards 5 instead of all - I do not remember. Actually there is no problem with this filter.
The only thing I changed with the user dict in beta 2 is: I created a new user dict in beta 2 and imported cards. I thought that beta 2 creates upgraded user dicts, because I did not find the upgrade button for this dict in manage dictionaries, even after I quit the program, but after a soft reset the upgrade button was there again, and after an upgrade of the user dict this dictionary was visible in compounds again. Therefore I think the user dict can only be seen in Compounds, if the user dict is upgraded.
Two more remarks: is it possible to change the font size in Card Info / Text/Categories? Can you use the full screen in Flashcards / Statistics?
Alexis
状元
Not sure if this is a bug or intentional, but I did a repitition-spaced flashcard session, then chose NO at the end. A few minutes later I decided I wanted to do some more, so I started a session and chose "yes". I was showed cards that I had already reviewed that day. Is it possible not to show a card that has already been reviewed that day?
HW60
状元
I found the history information in the XML file. For displaying the last 10 results would do. The information screen (with the Edit Card Edit Cats buttons) has a lot of free space both in Text and in Statistics on 480*800 screens - the new category selection screen really makes fun to select the right category.mikelove said:
Is there a documentation available for XML details (date items, number of records per card, ...)?
HW60 - that's true, the XML files are complicated to read in their raw form; we avoid inserting lots of extra whitespace to make them easier to read since we want to keep the files as small as possible. However, there are lots of XML viewer apps that might help with this, including several good free ones.
We did manage to add a History field for Beta 3 (though currently it's WM-only, Palm still has "Last 10" since there's less room), so we're just waiting a little while longer to see if any new bugs crop up that need to be fixed before we release that.
And yes, user dictionaries will only work in Compounds if they're upgraded - they'd be *incredibly* slow with the old indexing system (which turns out to have been the reason why they weren't supported before, I'd just forgotten about that). 2.0.8 Beta 1 and 2 create upgraded user dictionaries initially, and I just successfully tested creating / adding an entry to / searching for compounds in a new dictionary and it seems to work correctly - is there any chance you might have been mixing up an old and a new dictionary with the same name?
There's no way to change the font size in Card Info / Categories; is it too small or too big for you right now? Using the full screen for statistics really isn't necessary since we managed to squeeze History in there anyway, and in general we'd rather have as few of those auto-rearranging screens as possible since they're tricky to code and prone to lots of device-specific issues.
mfcb - OK; we will fix that correct / incorrect count issue for Beta 3 at least.
Alexis - that is intentional, yes; even though those cards were already reviewed, they're scheduled to come up sooner than any / most other cards, which is why they're coming up right away in the new session. If you'd like to skip them, though, just add a Card Filter "Time filter" for "last reviewed" "not within" "1" "days" (or "calendar days").
HW60 again - There's no documentation for the XML format yet, combination of programmer laziness and the fact that we're still trying to add support for a few important things it's lacking (like the ability to embed profile settings, so that you can really make a complete XML backup of your flashcard database - that may come in Beta 3, actually, since it'll help considerably with iPhone transfers).
We did manage to add a History field for Beta 3 (though currently it's WM-only, Palm still has "Last 10" since there's less room), so we're just waiting a little while longer to see if any new bugs crop up that need to be fixed before we release that.
And yes, user dictionaries will only work in Compounds if they're upgraded - they'd be *incredibly* slow with the old indexing system (which turns out to have been the reason why they weren't supported before, I'd just forgotten about that). 2.0.8 Beta 1 and 2 create upgraded user dictionaries initially, and I just successfully tested creating / adding an entry to / searching for compounds in a new dictionary and it seems to work correctly - is there any chance you might have been mixing up an old and a new dictionary with the same name?
There's no way to change the font size in Card Info / Categories; is it too small or too big for you right now? Using the full screen for statistics really isn't necessary since we managed to squeeze History in there anyway, and in general we'd rather have as few of those auto-rearranging screens as possible since they're tricky to code and prone to lots of device-specific issues.
mfcb - OK; we will fix that correct / incorrect count issue for Beta 3 at least.
Alexis - that is intentional, yes; even though those cards were already reviewed, they're scheduled to come up sooner than any / most other cards, which is why they're coming up right away in the new session. If you'd like to skip them, though, just add a Card Filter "Time filter" for "last reviewed" "not within" "1" "days" (or "calendar days").
HW60 again - There's no documentation for the XML format yet, combination of programmer laziness and the fact that we're still trying to add support for a few important things it's lacking (like the ability to embed profile settings, so that you can really make a complete XML backup of your flashcard database - that may come in Beta 3, actually, since it'll help considerably with iPhone transfers).
Small thing: The text of the new "session Complete" box does not fit exactly:
...
- continue reviewing cards that aren't due yet (Yes)
- stop reviewing (No)
- review those not-due-yet cards without extending their due dates (Cancel)
The text for "Yes" ist wrong.
"Yes" should be:
- review cards that aren't due yet without extending their due dates (Yes)
...
- continue reviewing cards that aren't due yet (Yes)
- stop reviewing (No)
- review those not-due-yet cards without extending their due dates (Cancel)
The text for "Yes" ist wrong.
"Yes" should be:
- review cards that aren't due yet without extending their due dates (Yes)
"Yes" actually is supposed to review cards that aren't due yet and extend their due dates; No stops reviewing, and Cancel reviews them without extending their due dates. Are you seeing different behavior on your system? The scores might not change even with "Yes" if you've enabled "only change once per day" in Advanced Settings / Scoring, but the "last reviewed" date should be updated so that the interval until the next review will still be extended.
HW60
状元
I had some problems creating a new user dict - Pleco proposed a name for the dict, then said, that this dict already exists. I finally deleted all user dicts and started again. Maybe the strange behaviour came from copying the user dict to a folder on my storage card - there I have a folder for Pleco dictionaries and fonts, and a folder for the flashcards (the program is in main storage\programs\Pleco). Now I created a new user dict in My documents without problems, and the upgrade button did not appear even after a soft reset. This time importing the same set of records (from exported flashcards) took approx. 1 second per card! I hope it is not due to my new 16GB SD card. So as a result: I cannot reproduce the upgrade story for user dicts.mikelove said:
There is another (?) strange thing with dictionaries in Compounds: one of my dictionaries is UNIHAN (p2unihan.pqb). When this dictionary is on a high rank in Manage Dictionaries, no other dictionary after UNIHAN is visible in Compounds. If I "down" UNIHAN to last position, all my dictinaries including 2 user dicts can be seen in Compounds. In the Main Screen I can see all dictionaries (and I took 一 (yi) for testing Compounds which is also included in UNIHAN). Actually this is not a problem, because now I leave UNIHAN in last position, but it takes some time to find out.
One more feature request: Compounds takes up to 2 seconds to appear depending on the size of the dictionary, therefore Character Info never starts with Compounds. I like compounds very much during flashcard testing, because when I do not know a character I can look in Compounds / user dict which words belong to this character. My user dict is an exact copy of all my flashcards. Therefore I find in Compounds all the words which I should know. My suggestion: either define the start tab of Character Info in Preferences, or even better keep the last open tab of Character Info for the next lookup of Character Info even if it is Compounds.
The font in Text / Categories is a little bit small, but a possible change is not too important. But I mean the Open Flashcards / Statistics, not the Card Info / Statistics. Here my favourite information Cards Learned always requires a scroll down of 1 line.mikelove said:
Not too important, but how do I convert the XML-dates (created, last reviewed,...) to understandable dates? Is it milliseconds since 1.1.1900 or something similar? Why is there sometimes one record per flashcard, sometimes 2 or 3 (also not too important)?mikelove said:
anchan42 - hmm... you're still using the dictionary entry remapping feature, right? ("Force defns from dicts") I just double-checked the code for that and I don't see any obvious sources of problems - could you give me some specific examples of flashcards whose definitions don't update correctly when you backtrack to them, along with the names of the dictionaries they normally link to and the current list of dictionaries selected for remapping? (and whether or not "Use custom card defns" / "as a fallback" are checked?)
HW60 - putting flashcard or user dictionary databases on an SD card is a Very Bad Idea - Windows Mobile's flash memory card drivers continue to be extremely unstable / unreliable, and in spite of SQLite's many protections against it, you're running a much higher risk of file corruption with the databases there than you are with them in internal memory.
User dictionary imports are slower in general with the new system - seems like a worthwhile tradeoff for the faster search performance, and after a few more revisions we may finally release some sort of desktop converter to use with large database imports. (if we don't just roll that functionality into Pleco Desktop, which is another possibility) Were you using "unsafe mode"? That should improve performance significantly on a slow SD card, since it means a lot of journaling / rollback data is kept only in internal memory instead of being written to the card.
I'm not seeing this Unihan compounds issue - could you let me know an exact character / Manage Dicts dictionary sort order for which you've observed it happening? We're considering a change in the next database format update that would make compounds searches in all databases extremely fast, so if/when we do that we'd then make Char Info return to Compounds just like any other tab.
Oh, and the dates in the XML files are standard Unix dates - number of seconds since 1/1/1970 00:00:00 UDT. http://www.epochconverter.com/ is a handy website for dealing with these. What do you mean by "one record per flashcard"?
HW60 - putting flashcard or user dictionary databases on an SD card is a Very Bad Idea - Windows Mobile's flash memory card drivers continue to be extremely unstable / unreliable, and in spite of SQLite's many protections against it, you're running a much higher risk of file corruption with the databases there than you are with them in internal memory.
User dictionary imports are slower in general with the new system - seems like a worthwhile tradeoff for the faster search performance, and after a few more revisions we may finally release some sort of desktop converter to use with large database imports. (if we don't just roll that functionality into Pleco Desktop, which is another possibility) Were you using "unsafe mode"? That should improve performance significantly on a slow SD card, since it means a lot of journaling / rollback data is kept only in internal memory instead of being written to the card.
I'm not seeing this Unihan compounds issue - could you let me know an exact character / Manage Dicts dictionary sort order for which you've observed it happening? We're considering a change in the next database format update that would make compounds searches in all databases extremely fast, so if/when we do that we'd then make Char Info return to Compounds just like any other tab.
Oh, and the dates in the XML files are standard Unix dates - number of seconds since 1/1/1970 00:00:00 UDT. http://www.epochconverter.com/ is a handy website for dealing with these. What do you mean by "one record per flashcard"?
mikelove said:
hi, i can confirm the unihan compounds issue, my order was ACE - ABC - OX - UNI - GF - ADS - HDD - CC - RAD - USR... after i put UNI to the end of the list i can have compounds from all dicts, not just the first 3... and its not just single characters, its all i checked, did not find one that did NOT show this issue...
BIG thanks to HW60 for finding this out, hehe, big improvement :lol:
HW60
状元
My user dict is always a copy of the flashcard db, but I will copy the flashcards to main storage because this is the valuable data. I make a copy of flashcards.pqb about once a week ...
There was no great difference in speed importing the flashcard export to to user dict either with or without unsafe mode, but your are right: the import happens only once.
Any dictionary sort order in Manage Dicts seems to lead to the same result: e.g with
ACED
Unihan
ABC
Oxford
etc.
I only get ACE in compounds. All dictionaries below Unihan and including Unihan are invisible in compounds. I marked all 4 ways of use for all dictionaries (show in menu, use in toolbar, ...). The more I move down Unihan, the more dictionaries are visible in compounds. I have ACED, ABC, Oxford, Tuttle, Guifan, Handedict, cc-cedict, adso, an upgraded User dict for classical chinese (from a forum member), radicals and 2 user dicts, both created within the last 2 weeks, one of them only for testing Beta 2. On the SD card in the Pleco folder are all dictionaries, the fonts Pleco Extra and ZYSong, the classical chinese user dictionary and a file Radical_Dict, last modified today, which I do not remember where it came from (there is a p2radicals file too). The other 2 user dicts are in MyDocuments.
When I exported my 6 test flashcards (to test Card Filter not correct in row) as XML file and opened this file in Excel, I expected 6 records or lines in Excel, but actually I saw about 14, some flashcards had 1 line/record, some flashcards had 2 or 3 lines/records. But if you include card history in Beta 3, I will probably no longer search for XML details.
There was no great difference in speed importing the flashcard export to to user dict either with or without unsafe mode, but your are right: the import happens only once.
Any dictionary sort order in Manage Dicts seems to lead to the same result: e.g with
ACED
Unihan
ABC
Oxford
etc.
I only get ACE in compounds. All dictionaries below Unihan and including Unihan are invisible in compounds. I marked all 4 ways of use for all dictionaries (show in menu, use in toolbar, ...). The more I move down Unihan, the more dictionaries are visible in compounds. I have ACED, ABC, Oxford, Tuttle, Guifan, Handedict, cc-cedict, adso, an upgraded User dict for classical chinese (from a forum member), radicals and 2 user dicts, both created within the last 2 weeks, one of them only for testing Beta 2. On the SD card in the Pleco folder are all dictionaries, the fonts Pleco Extra and ZYSong, the classical chinese user dictionary and a file Radical_Dict, last modified today, which I do not remember where it came from (there is a p2radicals file too). The other 2 user dicts are in MyDocuments.
When I exported my 6 test flashcards (to test Card Filter not correct in row) as XML file and opened this file in Excel, I expected 6 records or lines in Excel, but actually I saw about 14, some flashcards had 1 line/record, some flashcards had 2 or 3 lines/records. But if you include card history in Beta 3, I will probably no longer search for XML details.
mike,
the new option to integrate full-text search into regular English-Chinese / Chinese-English searches. should that work also on user-dicts?
i have a user dict with some example sentences (without pinyin, just hanzi/english) but when i do a search for "luggage" nothing comes up from my user dict (should have 4 occurrences)...
also no success with "#luggage", still ignores my user dict...
the dict is C-E, maybe i should swap that?
the new option to integrate full-text search into regular English-Chinese / Chinese-English searches. should that work also on user-dicts?
i have a user dict with some example sentences (without pinyin, just hanzi/english) but when i do a search for "luggage" nothing comes up from my user dict (should have 4 occurrences)...
also no success with "#luggage", still ignores my user dict...
the dict is C-E, maybe i should swap that?