2.0.4 Beta 4 Bug Report / Feedback Thread

[HW60]

I have more and more problems entering chinese characters during flashcard sessions (free-answer, show audio, ask for headword; hwr-auto-rec after: instant; HTC Touch HD). If I draw lines in normal speed, I usually get broken lines with a "scratching" feeling, therefore I have to draw the lines very slowly to get complete lines. Entering characters in the HWR box of the main screen is no problem. Do you have any idea what could be the reason and what I could do to get a "smooth" input also during flashcard sessions? To test I changed to left-handed input without improving the results; writing characters without "Filter radical results by character set" seems to improve the situation a little bit.

 

[marsch]

Re the new "/ and a third number in Card Info history screen - total number of times reviewed": If the numbers get too big, say 95-25/120, then the text is cut after the 95-, so I don't get to see the 25, let alone 120. Or maybe that's on purpose so I'm not constantly reminded I got it wrong so many times? :roll:

 

[mikelove]

HW60 - good point about Char Info compounds, we really want to accelerate those searches at some point but in the meantime it might at least be nice to have a tap-hold menu in the dictionary chooser there. With the difficult handwriting input, is there any chance your screen / screen protector might be getting scratched or damaged? That's the most common reason for input to start unexpectedly skipping, and since handwriting in flashcards is confined to such a small space it might be easy for that corner of the screen to get scratched with heavy use.

marsch - good point, need to try to find a few more pixels for that box.

 

[HW60]

HW60 - With the difficult handwriting input, is there any chance your screen / screen protector might be getting scratched or damaged? That's the most common reason for input to start unexpectedly skipping.

The HWR input box in the main screen has the same size, though it is not used as often as the flashcard box. I do not use a screen protector, and the screen is not too much scratched (the PDA is 6 months old). Actually the problems vary very much: half an hour ago I had a flashcard session without any problems, and this morning I was shortly before switching to self-scored because lots of broken lines. I always close all programs before I start, and I even tried a soft reset. My first impression was that during "repeat incorrect at end" the problem had disappeared, but then even the next full session was OK again. This session, however, started with a bug I reported Mar 27, 5:41 pm: for a 3 character word I only got one character box to fill. I have no idea if these 2 problems belong together. Since Mar 27 this was the 2nd time that the number of boxes did not match the number of characters.

 

[mikelove]

We still haven't managed to reproduce that bug with too few boxes appearing: which specific word(s) have you seen this happen with? Was it the same 3-character word in both cases?

On the handwriting issue, something might be sapping your system memory (or CPU) even if all programs are closed - if you go into Start / Settings / Memory when you're having flashcard handwriting problems versus when you're not, do you see a big difference in how much free memory your system has? Also, somewhat random - was there a big temperature difference between the times the recognizer was working well and when it wasn't? Even on your device, if you had, say, been using it for a while so it was running very warm one time and hadn't used it at all lately the other time. Or was it maybe plugged into an AC adapter one time and not plugged in another time?

 

[HW60]

The too few boxes bug always (i.e. about 5 or 6 times alltogether) happened with the first word of a flashcard session, always repetition-spaced, Automatic, free-answer, show audio (with earphone), ask for headword. It was always the second session of 2 sessions in a row, either both sessions alike or first session with free answer, show headword, ask for pronunciation, second session with show audio etc. I do not remember the words, but I think it is highly unlikely that two words were the same. I remember one 4 character word, one 3 character word and two 2 character words, but only the length, not the words. All my flashcards are custom flashcards.

I shall keep watching the conditions under which the HWR problems occur and report on the results.

 

[marsch]

Hi. is there a way to be notified when the next beta comes out? I have subscribed to the Thread and the Forum, but that way I get notified whenever there is any new msg and have to check each time whether that is about a new beta - which it usually isn't... Would be nice to have that feature (maybe by opening a thread specifically for that?) - or to know how to use it in case it exists already

 

[mikelove]

HW60 - that second-session bit is interesting; perhaps the UI is failing to update correctly somehow. But exiting and then resuming the session would bring back the right number of boxes, right?

marsch - good idea; see the newly-created http://plecoforums.com/viewtopic.php?f=11&t=2242.

 

[HW60]

HW60 - that second-session bit is interesting; perhaps the UI is failing to update correctly somehow. But exiting and then resuming the session would bring back the right number of boxes, right?

There is no need to do anything: the second flashcard (and all following) has its due number of boxes, and the first flashcard can be answered correctly even without boxes! I just enter the characters, and after the right number of characters is entered the visible boxes get green or red.

I changed my tests somewhat, and now I found something strange I do not know if it is inteded or a bug: Repetition spaced, Test subject selection alternating, self scored, test on reading and writing (reading: show headword, writing show pron + defn), repeat incorrect at end with loop and shuffle. During repetition of incorrect cards there is no alternating between reading and writing - only reading for all cards. As there are some cards I can read very well, but write very poorly, maybe it is a bug ...

 

[mikelove]

Reproduced and fixed; looks like we'd somehow stopped recording which mode each card was answered in in the session history. Thanks!

 

[mfcb]

mike, something which is a "feature" already very long time:
in the flashcard session start dialog on the page to select the session profile, when i focus the combobox to select the profile, the "new flashcard session" button gets the highlight, like it would have the focus. thats sometimes confusing... when other buttons in this dialog have the focus everything is ok...

 

[HW60]

HW60 - that second-session bit is interesting; perhaps the UI is failing to update correctly somehow. But exiting and then resuming the session would bring back the right number of boxes, right?

One last remark to this problem: I think it is an initial value problem, which sometimes occurs, when a program does not start from the beginning (= first flashcard session), but restarts (= second flashcard session, sometimes after a pause or even after having started and quit another program). When starting, the compiler or the programmer erases all values; restarting the programmer sometimes forgets to reinitialize a value, and then "any" value from memory is taken, which sometimes leads to an error, sometimes not. Here only the number for drawing boxes seems to be effected; for the program the length of the word is known, as I can enter characters in not visible boxes, and only when I entered the last character the result can be seen in the visible boxes. Probably Pleco even displays the entered characters, but some boxes do not show the "displayed" characters.

 

[poncin]

I have more and more problems entering chinese characters during flashcard sessions (free-answer, show audio, ask for headword; hwr-auto-rec after: instant; HTC Touch HD). If I draw lines in normal speed, I usually get broken lines with a "scratching" feeling, therefore I have to draw the lines very slowly to get complete lines. Entering characters in the HWR box of the main screen is no problem. Do you have any idea what could be the reason and what I could do to get a "smooth" input also during flashcard sessions? To test I changed to left-handed input without improving the results; writing characters without "Filter radical results by character set" seems to improve the situation a little bit.

Another case of of "ｂｒｏｋｅｎ　ｌｉｎｅｓ" handwriting.

context:　HTC　Ｄｉａｍｏｎｄ　/ Ｓｔａｋｓ　ＲＯＭ / QQｐｉｎｙｉｎ input method
1. Ｉｆ　Ｉ　ｌａｕｎｃｈ　ｐｌｅｃｏ，　ｍａｋｅ　ｔｈｅ　ＱＱｐｉｎｙｉｎ　ｋｅｙｂｏａｒｄ　ｖｉｓｉｂｌｅ　ａｎｄ　ｔｈｅｎ　ｗｉｔｃｈ　ｔｏ　ｔｈｅ　ｈａｎｄｗｒｉｔｉｎｇ　ｒｅｃｏｇｎｉｔｉｏｎ　ｓｃｒｅｅｎ，　ｔｈｅ　ｄｒａｗｉｎｇ　ｉｓ　ｖｅｒｙ　＂ｓｃｒａｔｃｈｙ＂. At this point, the keyboard, covers part of the screen.
Ｉｆ　Ｉ　ｍｉｎｉｍｉｚｅ　ｔｈｅ　ＱＱ ｋｅｙｂｏａｒｄ，　the drawing remains scratchy.
2.　Ｉｆ　ｂｅｆｏｒｅ　ｓｗｉｔｃｈｉｎｇ　ｔｏ　ｔｈｅ　ｈａｎｄｗｒｉｔｉｎｇ　ｒｅｃｏｇｎｉｔｉｏｎ　ｓｃｒｅｅｎ，　Ｉ　ｍｉｎｉｍｉｚｅ　ｔｈｅ　ＱＱ　Ｋｅｙｂｏａｒｄ，　ｔｈｅｎ　ｔｈｅ　ｗｒｉｔｉｎｇ　ｉｓ　ｖｅｒｙ　ｓｍｏｏｔｈ．　Ａｔ　ｔｈａｔ　ｐｏｉｎｔ，　ｉｆ　Ｉ　ｏｐｅｎ　ａｎｄ　ｃｌｏｓｅ　ｔｈｅ　ＱＱ　ｋｅｙｂｏａｒｄ，　ｔｈｅ　ｗｒｉｔｉｎｇ　ｂｅｇｉｎｓ　ｔｏ　ｂｅ　ｖｅｒｙ　ｓｃｒａｔｃｈｙ　ａｇａｉｎ．
３．　Ｉｆ　Ｉ　ｓｗｉｔｃｈ　ｂａｃｋ　ｔｏ　ｔｈｅ　ｍａｉｎ　ｓｃｒｅｅｎ　ａｎｄ　ｇｏｅｓ　ｂａｃｋ　ｔｏ　ｔｈｅ　ｈａｎｄｗｒｉｔｉｎｇ　ｓｃｒｅｅｎ，　ｔｈｅ　ｗｒｉｔｉｎｇ　ｉｓ　ｓｍｏｏｔｈ　ａｇａｉｎ．

－－
Ｆｒｅｄｅｒｉｃ　Ｐｏｎｃｉｎ

 

[poncin]

Another case of of "ｂｒｏｋｅｎ　ｌｉｎｅｓ" handwriting.

context:　HTC　Ｄｉａｍｏｎｄ　/ Ｓｔａｋｓ　ＲＯＭ / QQｐｉｎｙｉｎ input method
1. Ｉｆ　Ｉ　ｌａｕｎｃｈ　ｐｌｅｃｏ，　ｍａｋｅ　ｔｈｅ　ＱＱｐｉｎｙｉｎ　ｋｅｙｂｏａｒｄ　ｖｉｓｉｂｌｅ　ａｎｄ　ｔｈｅｎ　ｗｉｔｃｈ　ｔｏ　ｔｈｅ　ｈａｎｄｗｒｉｔｉｎｇ　ｒｅｃｏｇｎｉｔｉｏｎ　ｓｃｒｅｅｎ，　ｔｈｅ　ｄｒａｗｉｎｇ　ｉｓ　ｖｅｒｙ　＂ｓｃｒａｔｃｈｙ＂. At this point, the keyboard, covers part of the screen.
Ｉｆ　Ｉ　ｍｉｎｉｍｉｚｅ　ｔｈｅ　ＱＱ ｋｅｙｂｏａｒｄ，　the drawing remains scratchy.
2.　Ｉｆ　ｂｅｆｏｒｅ　ｓｗｉｔｃｈｉｎｇ　ｔｏ　ｔｈｅ　ｈａｎｄｗｒｉｔｉｎｇ　ｒｅｃｏｇｎｉｔｉｏｎ　ｓｃｒｅｅｎ，　Ｉ　ｍｉｎｉｍｉｚｅ　ｔｈｅ　ＱＱ　Ｋｅｙｂｏａｒｄ，　ｔｈｅｎ　ｔｈｅ　ｗｒｉｔｉｎｇ　ｉｓ　ｖｅｒｙ　ｓｍｏｏｔｈ．　Ａｔ　ｔｈａｔ　ｐｏｉｎｔ，　ｉｆ　Ｉ　ｏｐｅｎ　ａｎｄ　ｃｌｏｓｅ　ｔｈｅ　ＱＱ　ｋｅｙｂｏａｒｄ，　ｔｈｅ　ｗｒｉｔｉｎｇ　ｂｅｇｉｎｓ　ｔｏ　ｂｅ　ｖｅｒｙ　ｓｃｒａｔｃｈｙ　ａｇａｉｎ．
３．　Ｉｆ　Ｉ　ｓｗｉｔｃｈ　ｂａｃｋ　ｔｏ　ｔｈｅ　ｍａｉｎ　ｓｃｒｅｅｎ　ａｎｄ　ｇｏｅｓ　ｂａｃｋ　ｔｏ　ｔｈｅ　ｈａｎｄｗｒｉｔｉｎｇ　ｓｃｒｅｅｎ，　ｔｈｅ　ｗｒｉｔｉｎｇ　ｉｓ　ｓｍｏｏｔｈ　ａｇａｉｎ．
[/quote]

Ｄｉｇｇｉｎｇ　ｆｕｒｔｈｅｒ　．．．
Ｉ　ｃｈａｎｇｅｄ　ｔｈｅ　ｓｋｉｎ　ｏｆ　ｔｈｅ　ｋｅｙｂｏａｒｄ　ｉｎ　ｏｒｄｅｒ　ｔｏ　ｈａｖｅ　ａ　ｓｍａｌｌ　ｏｎｅ　ｃｏｖｅｒｉｎｇ　ａ　ｌｉｍｉｔｅｄ　ｐａｒｔ　ｏｆ　ｔｈｅ　ｓｃｒｅｅｎ．　Ｉｎ　ｔｈｉｓ　ｃａｓｅ，　ｔｈｅ　ｋｅｙｂｏａｒｄ　ｗｉｌｌ　ｎｏｔ　ｏｖｅｒｌａｐ　ｔｈｅ　ｂｏｘ　ｗｈｅｒｅ　ｔｈｅ　ｈａｎｄｗｒｉｔｉｎｇ　ｒｅｃｏｇｎｉｔｉｏｎ　ｉｓ　ｐｅｒｆｏｒｍｅｄ．　Ｉｎ　ｓｕｃｈ　ｃｏｎｄｉｔｉｏｎｓ，ｔｈｅ　ｐｒｅｓｅｎｃｅ　ｏｆ　ｔｈｅ　ｋｅｙｂｏａｒｄ　ｏｎ　ｔｈｅ　ｓｃｒｅｅｎ　ｈａｓ　ｎｏ　ｉｍｐａｃｔ　（ｗｒｉｔｉｎｇ　ｉｓ　ｓｍｏｏｔｈ）．　Ｓｗｉｔｃｈｉｎｇ　ｉｔ　ｏｎ／ｏｆｆ　ｉｓ　ｆｉｎｅ．

 

[HW60]

I still have no idea under which conditions the HWR seems to scratch the characters. It seems that during "Repeat incorrect at end" the writing is smooth, while it was scratchy during the normal session. Maybe that holding the PDA (HTC Touch HD) in one hand (instead of placing it on a table) and writing with the other has a negative impact, but these are only first impressions.

 

[HW60]

In repetition spaced, free-answer, show audio, ask for headword the character input box is erased, when a character from the selection box is selected. Sometimes, especially when I made a mistake writing a character and Pleco shows a red box as the result of this mistake, I write the character again into the input box to learn writing. When I switch to the next flashcard by hitting the 答-button, the input box is not erased.

I think the moment to erase the character from the input box should be different in both cases: it may be interesting to see my written character in the input box, when Pleco shows that I made a mistake. And there seems to be no reason to keep the written character when switching to the next flashcard - the probability that the next word starts with the character written in the input box can be neglected.

Of course it is not necessary to abuse the input box for learning purposes ...

 

[mikelove]

mfcb - we'll take a look at that one, might have something to do with some changes we made explicitly to ensure that the New Session button would be highlighted whenever that screen comes up.

HW60 - that's my sense of it too, but the problem is that we don't actually store the number of boxes anywhere in the flashcard system code; the program fetches the headword text (from the same place where it gets the rest of the card's definition, and at the same time) and immediately puts the appropriate number of boxes up. So it has to be some sort of a UI problem, a message firing off too early / too late that causes the number of boxes to be reset without the card even being loaded yet.

poncin - must be some sort of graphical compositing problem; the system's taking a while to blend the keyboard foreground with the Pleco window behind it. (we've seen a TON of these issues on iPhone, actually) Thanks.

HW60 again - good point, we might as well clear the box in that case.

 

[Wan]

Hi folks,

this is not a bug. Actually, I'm not quite sure whether this is anything at all: From the last program update, the startup time of Pleco has at least doubled. Everything else, every function I've tried, works fine, though. This may be due to older hardware (Touch Diamond shouldn't pose a problem?!) or because I licensed two more dictionaries and it's just a lot more to load? Maybe it's the program settings, but I don't think so. One other thing: When recognizing my key file, Pleco tells me not all of the products I've licensed are found. I'm sure all dictionaries are there, maybe it's the voices (I completely deleted them).

Greetings,

Wan

 

[mikelove]

It might be the stuff we've done to minimize the number of files Pleco has open at once, actually (which in general seems to be doing very good things for system stability) - so this didn't happen in Beta 3 but only started in Beta 4, or it didn't happen in 2.0.3 but happened once you started using the 2.0.8/2.0.4 (boy that's confusing, we really blew it on that one) betas?

 

[Wan]

so this didn't happen in Beta 3 but only started in Beta 4, or it didn't happen in 2.0.3 but happened once you started using the 2.0.8/2.0.4 (boy that's confusing, we really blew it on that one) betas?

Now that you mention it, I believe it started in the previous beta [Beta 2.0.3].