llammamama
探花
I've sent a copy of my flashcard database along with my scorefile (which turned out to be an empty file) in an email to your mikelove (at) pleco acount. I have some more details of my problem in the email.
Beta 5 Bug Report / Feedback Thread
- Thread starter mikelove
- Start date
ldolse - even getting people to tap on buttons is sometimes a stretch; we still get tons of people who don't know that the zhong/ying button switches between English and Chinese and can't figure out how to access the E-C dictionaries. (not all that unreasonable when you consider that it's at the top left corner of the screen and could easily be a program icon / logo of some kind) Anyway we'll make sure those bugs are fixed with the "keep on clear" option in the next beta.
grabolus - glad to hear it's working well. The stroke order diagram code was moved over to ARM for the latest build (when it does work you might notice that it draws a lot faster, and in the test mode now has a cool feature where if you hold the stylus down it fills in the stroke it thinks you're about to select in blue so you can move it if you've accidentally hit the wrong one) so that's why people are having all of these new problems with it. Has instant access worked for you at other times or does it always reset like that? (we've already gotten data files for HanDeDict, with flashcard-friendly UIDs even, but have been too bogged down fixing buts to deal with them)
sfrrr - glad to hear it, short report or no
Do you remember if you had any other controls visible (input palette, etc) when the input box lost the keyboard focus? All of the controls on the main screen are supposed to send the focus back to the input field whenever they receive a keystroke (Windows Mobile will sometimes semi-arbitrarily take the focus away), but one of them might not be doing it correctly. The toolbar only supports 10 icons because the minimum spacing for them is 24 pixels (or 48 on VGA) - the drop box doesn't stop at 10 because some people might prefer to keep the same configuration on portrait/landscape screens but have extra buttons appear on the end in landscape. SQLite on Windows Mobile was actually developed entirely by them, all we did was stick the source file into our application and build it, so if you run into any data corruption issues now you have a much larger circle of people you can blame
ipsi - both magnification and SC/TC can be configured in the Display panel in Advanced settings. Setting score-based field selection in the General tab (sorry the flashcard documentation isn't ready yet but we've been spending all day pounding away on this blasted singlehandedly-turning-me-into-a-Palm-hater TX bug) will let you specify which fields to test based on ranges of scores. Not sure about that lost data, might be the unorthodox import format or might be a lesser version of the same problem plaguing TXes.
weight/Fleminator - you're both quite sure you have enough file storage space for these exported files, right? They can get pretty big, 5 MB or more. Also, could you each give me a list of all of the other third-party software you've installed on your TX? Since we have yet to observe this problem with the converter on a single system here, TX or otherwise, I'm starting to think it might be some sort of a software conflict.
llammamama - thanks, hopefully we'll get to the bottom of that soon.
grabolus - glad to hear it's working well. The stroke order diagram code was moved over to ARM for the latest build (when it does work you might notice that it draws a lot faster, and in the test mode now has a cool feature where if you hold the stylus down it fills in the stroke it thinks you're about to select in blue so you can move it if you've accidentally hit the wrong one) so that's why people are having all of these new problems with it. Has instant access worked for you at other times or does it always reset like that? (we've already gotten data files for HanDeDict, with flashcard-friendly UIDs even, but have been too bogged down fixing buts to deal with them)
sfrrr - glad to hear it, short report or no
Do you remember if you had any other controls visible (input palette, etc) when the input box lost the keyboard focus? All of the controls on the main screen are supposed to send the focus back to the input field whenever they receive a keystroke (Windows Mobile will sometimes semi-arbitrarily take the focus away), but one of them might not be doing it correctly. The toolbar only supports 10 icons because the minimum spacing for them is 24 pixels (or 48 on VGA) - the drop box doesn't stop at 10 because some people might prefer to keep the same configuration on portrait/landscape screens but have extra buttons appear on the end in landscape. SQLite on Windows Mobile was actually developed entirely by them, all we did was stick the source file into our application and build it, so if you run into any data corruption issues now you have a much larger circle of people you can blame ipsi - both magnification and SC/TC can be configured in the Display panel in Advanced settings. Setting score-based field selection in the General tab (sorry the flashcard documentation isn't ready yet but we've been spending all day pounding away on this blasted singlehandedly-turning-me-into-a-Palm-hater TX bug) will let you specify which fields to test based on ranges of scores. Not sure about that lost data, might be the unorthodox import format or might be a lesser version of the same problem plaguing TXes.
weight/Fleminator - you're both quite sure you have enough file storage space for these exported files, right? They can get pretty big, 5 MB or more. Also, could you each give me a list of all of the other third-party software you've installed on your TX? Since we have yet to observe this problem with the converter on a single system here, TX or otherwise, I'm starting to think it might be some sort of a software conflict.
llammamama - thanks, hopefully we'll get to the bottom of that soon.
Ah! Right, didn't associate Subject Selection with the fields to be displayed. Not a big deal.
Also, I typically associate Cancel with clearing any changed data, which is not the case on the Flashcard screens. I'd suggest either changing it to Close, or adding a Save button
.
Getting occasional crashes on exiting, no idea why. Maybe flashcard related? Couldn't say.
Can't force simplified/traditional display. Button is ticked but nothing changes. Preferences for Magnified and such are also not working. They don't stay ticked.
I think the problem with the data was just deleting the wrong scorefile. It doesn't help the Profiles/Scorefiles screen is so slow and generally unresponsive.
Also, when telling a new profile to use the default scorefile it ignores you and creates a new one instead.
That's all I've got for now. I like the spinning thing when starting a flashcard session that tells you its doing something.
Also, I typically associate Cancel with clearing any changed data, which is not the case on the Flashcard screens. I'd suggest either changing it to Close, or adding a Save button
.Getting occasional crashes on exiting, no idea why. Maybe flashcard related? Couldn't say.
Can't force simplified/traditional display. Button is ticked but nothing changes. Preferences for Magnified and such are also not working. They don't stay ticked.
I think the problem with the data was just deleting the wrong scorefile. It doesn't help the Profiles/Scorefiles screen is so slow and generally unresponsive.
Also, when telling a new profile to use the default scorefile it ignores you and creates a new one instead.
That's all I've got for now. I like the spinning thing when starting a flashcard session that tells you its doing something
.sfrrr said:
Sandra, I'm using the Palm version, but there's almost no case I've come across where the audio doesn't say anything - either by using a recording of a multi-syllable word (e.g., 你好 is definitely present, as are lots of others) or by speaking recordings of individual syllables. So you should be getting at least *something* in almost all cases.
Can I suggest: check in preferences (under External, I think) that the "Skip words w/o multi-syll audio" option isn't turned on, and check that you've installed all the audio files in the right place (e.g. the same place as your dictionary files). Also check (in Help -> Ordering, I think) that the Audio feature is unlocked.
I had a problem with audio when I first installed, and it turns out that my audio files were corrupted when copied to my SD card because I ran out of space on the card when I was copying. To fix that I eventually had to delete and replace the files, and re-install Pleco 2 and re-apply the key file.
Another Import error, character in the flashcard file is 前, the import file entry has the Pinyin qian2
I get 6 matches in Adso, five of which are fine. The sixth is the problem. Adso headword is structured as 前来[來]qian2lai2. I'm guessing it matched based on the fact that the traditional headword is a single character, but seeing as this isn't even the character in the flashcard file this shouldn't be matching at all.
I get 6 matches in Adso, five of which are fine. The sixth is the problem. Adso headword is structured as 前来[來]qian2lai2. I'm guessing it matched based on the fact that the traditional headword is a single character, but seeing as this isn't even the character in the flashcard file this shouldn't be matching at all.
Allowing ambiguous matches during import is also preventing the search to go to the next dictionary to find a correct match.
Example is 着 zhe5 is in the import file with Adso as the first dictionary. Apparently this character doesn't exist in Adso standalone, But there are two two character words starting with this character. Rather than going to the next dictionary to find a better match the import routine forces the user to choose an incorrect match or reject the card.
Example is 着 zhe5 is in the import file with Adso as the first dictionary. Apparently this character doesn't exist in Adso standalone, But there are two two character words starting with this character. Rather than going to the next dictionary to find a better match the import routine forces the user to choose an incorrect match or reject the card.
Hi Mike,
I don't understand what you have written about the Palm TX. I am specially interested in the effect not in the technical reason. Does it mean there is sometimes a problem with import? Does it mean there is always a problem with import? Does it mean the import doesn't show mistake but didn't catch everything correctly? Does it mean everything is unstable on Palm TX?
So this are my first impressions on a Palm TX with beta5:
Import of about 8500 entires takes long but seem to work. I imported it over night because it took about 3 hours. Don't know if there were any problems because I didn't looked at it the whole time. When I looked today morning the palm was up and it seems that import worked.
1) When I copy something from dictionary entry to the clipboard (use /c) and try to open my dictionary software SlovoEd to translate English into German (this will be done automatically) it works a few times and then leads to a soft reset. It worked without any problem under PD 1.3 for more than 6 month. So probably a PD 20b5 problem. I just do the copy and then see the softreboot
2)
a)I tried the flashcards. It contains maybe 50 categories. I tried 5 different categories always containing about 20 to 50 cards. I used the automatically audio play. I never reached 20 cards because before it leads o a soft reset always just before playing the audio.
b) My original cards sometimes contain the Chinese word followed by the measure word in brackts: "aaa (bbb)". The brackets and its content wasn't exported or wasn't imported. It was lost anywhere during export and import.
c) Sometimes the entry contains normal numbers. This leads to a softreset for playing the audio.
d) I don't undertand how I can manage the cards. The old PD 1.3 flascards were much easier to handle. I want to see which cards belong to a category. I want to edit them. I want to move one card to another category. When I go to "Manage" I just see one category. Where are the 50 others?. When I go to "Manage Flashcards" I see all my cards but I don't know how to edit any or ho to move one to another category? When I go to "Manage Categories" I can change its name but how can I move a card to another? Don't have any idea how to handle this.
3) How can I do fulltext search? I would like to use it to look for a German word and find the right entry in the user dict.
4) I cannot look at the user dictionary. When I try to move to the user dict it always immediately switches back to ABC dictionary.
5) From time to time it soft reboots. Don't know what I did that could be the reason.
So it's still not seriously usable just for do some testing. I was very concerned that you not plan to support Palm in the future. I am very unhappy that you wrote there were problems with the TX that are difficult to fix.
From my point of view the progress from beta4 to beta5 is very few for the user, maybe very much for the code. I don't have the feeling that this is near to a usable system and I am afraid that Palm don't have enough of your interest.
I don't understand the technical reasons of the Palm TX problems. I just want to use it. I don't need the complicated new flashcard features that I cannot manage. I don't need a new database structure using SQlite if it causes problems and is so slow during import. Th old import was very fast. I wish there would be a version of PD 1.3 just having audio in flascards and fulltext search but as stable and as fast and as easy as before.
For me it means I am afraid I have to stay at good old PD 1.3 also in future. It's a pity that I will miss fulltext search and audio but it works nearly perfectly.
Nevertheless I hope there will be a mirracle and everyhing could be fixed to a stable and usable system. But I guess it will take minimum many more months.
Daniel
I don't understand what you have written about the Palm TX. I am specially interested in the effect not in the technical reason. Does it mean there is sometimes a problem with import? Does it mean there is always a problem with import? Does it mean the import doesn't show mistake but didn't catch everything correctly? Does it mean everything is unstable on Palm TX?
So this are my first impressions on a Palm TX with beta5:
Import of about 8500 entires takes long but seem to work. I imported it over night because it took about 3 hours. Don't know if there were any problems because I didn't looked at it the whole time. When I looked today morning the palm was up and it seems that import worked.
1) When I copy something from dictionary entry to the clipboard (use /c) and try to open my dictionary software SlovoEd to translate English into German (this will be done automatically) it works a few times and then leads to a soft reset. It worked without any problem under PD 1.3 for more than 6 month. So probably a PD 20b5 problem. I just do the copy and then see the softreboot
2)
a)I tried the flashcards. It contains maybe 50 categories. I tried 5 different categories always containing about 20 to 50 cards. I used the automatically audio play. I never reached 20 cards because before it leads o a soft reset always just before playing the audio.
b) My original cards sometimes contain the Chinese word followed by the measure word in brackts: "aaa (bbb)". The brackets and its content wasn't exported or wasn't imported. It was lost anywhere during export and import.
c) Sometimes the entry contains normal numbers. This leads to a softreset for playing the audio.
d) I don't undertand how I can manage the cards. The old PD 1.3 flascards were much easier to handle. I want to see which cards belong to a category. I want to edit them. I want to move one card to another category. When I go to "Manage" I just see one category. Where are the 50 others?. When I go to "Manage Flashcards" I see all my cards but I don't know how to edit any or ho to move one to another category? When I go to "Manage Categories" I can change its name but how can I move a card to another? Don't have any idea how to handle this.
3) How can I do fulltext search? I would like to use it to look for a German word and find the right entry in the user dict.
4) I cannot look at the user dictionary. When I try to move to the user dict it always immediately switches back to ABC dictionary.
5) From time to time it soft reboots. Don't know what I did that could be the reason.
So it's still not seriously usable just for do some testing. I was very concerned that you not plan to support Palm in the future. I am very unhappy that you wrote there were problems with the TX that are difficult to fix.
From my point of view the progress from beta4 to beta5 is very few for the user, maybe very much for the code. I don't have the feeling that this is near to a usable system and I am afraid that Palm don't have enough of your interest.
I don't understand the technical reasons of the Palm TX problems. I just want to use it. I don't need the complicated new flashcard features that I cannot manage. I don't need a new database structure using SQlite if it causes problems and is so slow during import. Th old import was very fast. I wish there would be a version of PD 1.3 just having audio in flascards and fulltext search but as stable and as fast and as easy as before.
For me it means I am afraid I have to stay at good old PD 1.3 also in future. It's a pity that I will miss fulltext search and audio but it works nearly perfectly.
Nevertheless I hope there will be a mirracle and everyhing could be fixed to a stable and usable system. But I guess it will take minimum many more months.
Daniel
daniel123 -- I'm using Beta 5 on my Palm Treo 680, and I'm getting some of the same problems you are. I'm replying in order to try to help you where I can, and also in order to provide more information to Mike so he can fix the problems.
I am also experiencing crashes when using audio in flash cards. For me, the problem usually occurs after about 60-80 cards. The device soft-resets at the time it would normally play the audio. Sometimes, it stops playing the audio for one or two, and then crashes on the third or fourth.
Mike, I also get a similar effect if I just go into a normal dictionary item and play the audio repeatedly. In one case, I looked up 光 in ABC, tapped on the play audio toolbar button, waited for the audio to finish, then tapped again. After repeating this approximately 70 times, the audio stopped playing. After repeating a few more times, the device soft reset.
It took me a little while to work this out, too, and I agree that the UI needs to be improved. What you need to do is go to "Manage Flashcards" now, instead of "Manage Categories". From there, tap the drop-down list at the top-left and choose "category". Then, on the right-hand side, choose the category that contains the cards you are looking for. Next click the 搜 character (sou1 - search) at the right-hand side. All the cards in that category will be displayed.
To put the flashcard in another category, select the target category from the drop-down at the bottom, then click the "+" icon near the bottom-left. Remember that in Pleco 2, cards can be members of multiple categories at the same time. So, if you want to move a card from one category to another, first you must copy it to the new category, then remove it from the old. To remove a card from a category, select the category you want to remove it from using the list at the bottom of the screen (next to the +/- buttons), then click the "-" button.
Fulltext search is supposed to turn on automatically if a normal match can't be found. For example, if I type "ikea", it automatically switches to fulltext search, and then finds 宜家 in my ADSO dictionary, because it's not present in my Oxford English-Chinese dictionary. However, to force a full-text search, you're supposed to be able to prefix the search with a hash (#) character... e.g. type "#ikea".
However, if I do this, it doesn't seem to search my user dictionary. Mike, perhaps you can explain why.
daniel123 said:
I am also experiencing crashes when using audio in flash cards. For me, the problem usually occurs after about 60-80 cards. The device soft-resets at the time it would normally play the audio. Sometimes, it stops playing the audio for one or two, and then crashes on the third or fourth.
Mike, I also get a similar effect if I just go into a normal dictionary item and play the audio repeatedly. In one case, I looked up 光 in ABC, tapped on the play audio toolbar button, waited for the audio to finish, then tapped again. After repeating this approximately 70 times, the audio stopped playing. After repeating a few more times, the device soft reset.
daniel123 said:
It took me a little while to work this out, too, and I agree that the UI needs to be improved. What you need to do is go to "Manage Flashcards" now, instead of "Manage Categories". From there, tap the drop-down list at the top-left and choose "category". Then, on the right-hand side, choose the category that contains the cards you are looking for. Next click the 搜 character (sou1 - search) at the right-hand side. All the cards in that category will be displayed.
To put the flashcard in another category, select the target category from the drop-down at the bottom, then click the "+" icon near the bottom-left. Remember that in Pleco 2, cards can be members of multiple categories at the same time. So, if you want to move a card from one category to another, first you must copy it to the new category, then remove it from the old. To remove a card from a category, select the category you want to remove it from using the list at the bottom of the screen (next to the +/- buttons), then click the "-" button.
daniel123 said:
Fulltext search is supposed to turn on automatically if a normal match can't be found. For example, if I type "ikea", it automatically switches to fulltext search, and then finds 宜家 in my ADSO dictionary, because it's not present in my Oxford English-Chinese dictionary. However, to force a full-text search, you're supposed to be able to prefix the search with a hash (#) character... e.g. type "#ikea".
However, if I do this, it doesn't seem to search my user dictionary. Mike, perhaps you can explain why.
westmeadboy
榜眼
Well, I didn't know about that tap and hold feature and I never came across it accidentally, so never got myself into a situation I couldnt get out of. But....
... now that you mentioned it, I tried it out - I hold down the stylus on the Toggle Input and select "Bottom" and, sure enough, the Input panel comes up at the bottom of the screen.
If I do the same, selecting Bottom (which is now ticked), then the Input panel disappears (and Bottom remains ticked) - actually I would expect it to do nothing (but not a problem).
Now if I select Full, then the full Input panel comes up and Full is ticked (no problem).
Everything ok so far.
However, if I select Bottom so that the Input panel comes up at the Bottom (bottom is now ticked)
Next I select Full, and this brings up the full Input panel (ok so far)
Now tap "ok" to go back to the main screen
But the Input panel still appears at the bottom and if I tap and hold the Toggle Input button I see "Bottom" is ticked.
This seems like a bug, although not a very serious one, admittedly
To summarize, with the Input panel displayed at the bottom try selecting Full from the dropdown and then tap "ok" - its still at the bottom.
Workaround - There is an easy way to get rid of the "bottom" Input panel and that is to reselect "Bottom" in the drop down.
The behaviour of the Toggle List button drop down is not the same. It seems to work intuitively.
... now that you mentioned it, I tried it out - I hold down the stylus on the Toggle Input and select "Bottom" and, sure enough, the Input panel comes up at the bottom of the screen.
If I do the same, selecting Bottom (which is now ticked), then the Input panel disappears (and Bottom remains ticked) - actually I would expect it to do nothing (but not a problem).
Now if I select Full, then the full Input panel comes up and Full is ticked (no problem).
Everything ok so far.
However, if I select Bottom so that the Input panel comes up at the Bottom (bottom is now ticked)
Next I select Full, and this brings up the full Input panel (ok so far)
Now tap "ok" to go back to the main screen
But the Input panel still appears at the bottom and if I tap and hold the Toggle Input button I see "Bottom" is ticked.
This seems like a bug, although not a very serious one, admittedly
To summarize, with the Input panel displayed at the bottom try selecting Full from the dropdown and then tap "ok" - its still at the bottom.
Workaround - There is an easy way to get rid of the "bottom" Input panel and that is to reselect "Bottom" in the drop down.
The behaviour of the Toggle List button drop down is not the same. It seems to work intuitively.
Hi Mike,
concerning your question: Instant Access worked once from a memo. It seemed not to work on a sms.
Today I lost ALL sms in my inbox(Messaging). I was using beta 5 and I was just on the character info tab when a sms arrived. I clicked on goto sms and a reset occured. Then inbox was empty and the categories of many applications were not set.
Fortunately I am doing a daily backup...
Another issue: After rearranging Input palette and playing around with full screen list it can happen that I get a strange screen update: First it draws some lines and the icons, after some seconds the full screen shows a list of entries, after some seconds a normal but slightly corrupt screen shows up etc. After a minute the screen doesn't change any more and I can go to preferences to go back to standard setting. Then everything is ok again...
Thanks
concerning your question: Instant Access worked once from a memo. It seemed not to work on a sms.
Today I lost ALL sms in my inbox(Messaging). I was using beta 5 and I was just on the character info tab when a sms arrived. I clicked on goto sms and a reset occured. Then inbox was empty and the categories of many applications were not set.
Fortunately I am doing a daily backup...
Another issue: After rearranging Input palette and playing around with full screen list it can happen that I get a strange screen update: First it draws some lines and the icons, after some seconds the full screen shows a list of entries, after some seconds a normal but slightly corrupt screen shows up etc. After a minute the screen doesn't change any more and I can go to preferences to go back to standard setting. Then everything is ok again...
Thanks
Hey Mike,
Really liking this 2.0 - here's a few possible bugs I found:
Soft resets:
~When I try to exit Pleco from Character info screen by pressing Palm Application/Home button, I get a soft reset [mentioned by Alexis earlier also]
~Flashcard session: Settings- repetition spaced, any test type, day type=hours (didn't try calendar), scoring type=automatic. After going through all cards in a category, if I tried to repeat the session it soft reset every time after that. I would expect to see a message "no cards to test" or something to that effect. This happened for every category I tried to repeat a session on. I tried adding a card between sessions to see if at least the one new card would test OK and it did with no reset. After waiting one day I could test through the cards again, but only one time per category as before - the second and following attempts caused soft resets. No problems at all on Freq adjust.
Sound files:
~After listening to several dozen audios, it will stop playing and then soft reset. (happens consistently)
~I've heard several audios files (woman's voice) with quite a bit of static also - not sure if you're interested in knowing, but one example is "li3"
Full text search:
~Read your comments about the full text search and I understand how it works, but having a problem when "Always search dicts in order" box is checked (Prefs/results). When searching English, only the NWP dict results will display and it won't switch to a different dict. If I uncheck "Use in search" (under manage dicts) for NWP, then only ABC results will come up. If I uncheck for ABC, only Oxford comes up and so on, but I can only get one dict to show any results. I think seaching for Chinese was similar but different order of dicts and in both cases it wasn't even in the order that was specified under mangage dicts. When the "Always search dicts in order" box is unchecked there are no problems and I can scroll through all the dicts that have results.
Thanks, Mator
Really liking this 2.0 - here's a few possible bugs I found:
Soft resets:
~When I try to exit Pleco from Character info screen by pressing Palm Application/Home button, I get a soft reset [mentioned by Alexis earlier also]
~Flashcard session: Settings- repetition spaced, any test type, day type=hours (didn't try calendar), scoring type=automatic. After going through all cards in a category, if I tried to repeat the session it soft reset every time after that. I would expect to see a message "no cards to test" or something to that effect. This happened for every category I tried to repeat a session on. I tried adding a card between sessions to see if at least the one new card would test OK and it did with no reset. After waiting one day I could test through the cards again, but only one time per category as before - the second and following attempts caused soft resets. No problems at all on Freq adjust.
Sound files:
~After listening to several dozen audios, it will stop playing and then soft reset. (happens consistently)
~I've heard several audios files (woman's voice) with quite a bit of static also - not sure if you're interested in knowing, but one example is "li3"
Full text search:
~Read your comments about the full text search and I understand how it works, but having a problem when "Always search dicts in order" box is checked (Prefs/results). When searching English, only the NWP dict results will display and it won't switch to a different dict. If I uncheck "Use in search" (under manage dicts) for NWP, then only ABC results will come up. If I uncheck for ABC, only Oxford comes up and so on, but I can only get one dict to show any results. I think seaching for Chinese was similar but different order of dicts and in both cases it wasn't even in the order that was specified under mangage dicts. When the "Always search dicts in order" box is unchecked there are no problems and I can scroll through all the dicts that have results.
Thanks, Mator
ipsi - it actually did act like a Cancel button originally but for various reasons we had to stop it from being so - we'll either bring that back or rename it in the finished version. Crash on exit certainly might be a flashcard database issue, we're really pounding away on that now and have already managed to make flashcard databases on Palm both faster and, oddly enough, smaller, so hopefully soon we'll have all of those problems cleared up. Profile/Scorefile managers don't cache their data like the Categories one yet, another feature we need to get around to adding at some point.
ldolse - we really didn't do any testing for this beta with Adso imports (didn't even get Adso converted until a day or two before we released it), so I'm not surprised to see a few problems like this - the other dictionaries are all very meticulous about making sure every character in a headword has a single-character entry, but obviously since Adso's not designed as a printed dictionary and hence don't need to make characters easy to find for radical table lookups it's not as big a concern for them. But yeah, we probably should add a dictionary switch button to that screen.
daniel123 - you're darn right we're not that committed to Palm, frankly the biggest mistake I've made in the history of running Pleco was deciding to put 2.0 on Palm OS; if we'd made it WM-only you'd all be using an uber-stable Pleco 2.1 on your iPAQs and admiring screenshots of the about-to-be-released iPhone version. We're losing large amounts of money for every day we spend working on Palm rather than some other platform with better business prospects. We've made the mistake of telling people for the last few years that they can still go ahead and buy a Palm to run Pleco (this has now been corrected on the PDA recommendations page incidentally), and for the sake of them and our large base of existing users we're doing all we reasonably can to get it working well on Palm, but at some point we have to move on to other projects even if it's not quite as stable and fast as the old version.
So we'll release what we can on Palm, we'll keep offering/supporting 1.0.3 for as long as anybody's interested in it (if there's enough demand we might even consider backporting a couple of features like audio and the new dictionaries to the old system), but if we lose a few customers or a few upgrade purchases because it's less than it could be on Palm, that's a reasonable price to pay for finally getting out of the Development Cycle Which Will Not End and moving on to the many other projects which we really ought to be working on. We made a lot of mistakes in planning 2.0, but even worse than those mistakes would be destroying Pleco's future in an attempt to fix every last one of them - at some point the best strategy is to apologize for what you can't fix, take your lumps and move on.
Re your specific issues, we'll see what we can find with this SlovoEd bug - hopefully it'll prove to be something minor / easy to find (as most such conflicts are). Audio's still a bit unstable after large numbers of plays, as a few other people have mentioned. The bracket-stripping is a bug in the flashcard creation function - it's supposed to remove rounded brackets from Pleco-supplied dictionaries like the ABC et al, since they're always just character variants like (兒), but it's not supposed to do so with user-created dictionary headwords.
The flashcard manager is a bit of a jump from 1.0 but I think once you get the hang of it it's really just as usable. Given the large negative response we've gotten to it, though, we are considering bringing back the old interface from 1.0 too (as an "Organize Flashcards" screen) - there'd still only be one copy of the flashcard in the system (unless you told it to be otherwise) but you could now move / copy a link to it between categories just like in the old version.
Full-text search isn't supported in user dictionaries, unfortunately - just not technically feasible on Palm to generate/update a full-text dictionary index on the fly. (would be on PPC but that'll have to wait for a later version) Not sure why it could be disappearing, might be the file corruption issues we're seeing elsewhere with the TX - hopefully we'll get to the bottom of those soon. (for what it's worth, I essentially haven't left my computer except to sleep in the last 3 days working on almost nothing but this TX bug)
sych - the audio system still isn't quite perfect, yeah - we made it a lot better since Beta 4 but there are still a few occasional memory leaks we're trying to dig up.
westmeadboy - that's actually what it's supposed to do, oddly enough (and what it's done for several versions now) - basically it works that way so that you can use it both to open the fullscreen input dialog and to open/close or change the position of the palette version. (since you might potentially want a larger screen for some inputs but be OK with the palette for others) Since we think most people will now be configuring the input palette using Preferences (much more sensible/intuitive anyway) we thought we might as well leave it working the way it already did in 1.0 for the sake of people who'd gotten used to it. (a logic we probably could have used in a few other places in 2.0 in hindsight)
grabolous - yikes, that's not good; the Messaging data must not have been saved out to the Treo's flash memory correctly when that reset happened. The 650 had a very early version of Palm's flash memory storage system so problems like that aren't as rare as we'd like them to be...
Mator - yeah, we've noticed this in a few places actually - a quirk of Palm OS software design is that the OS sends you through a completely different set of code when you exit a dialog box deliberately (thanks to a button press) than when it forces you to exit, so one of the many little debugging tasks we still haven't gotten around to yet is trying every one of those dialogs with a forced exit and making sure it doesn't crash. Not sure why it's crashing instead of gracefully exiting those no-card sessions but we've gotten reports of that even on PPC so it seems pretty widespread. The static in the woman audio isn't something we can do much about, unfortunately, the original recordings have the same static. Hadn't noticed that bug with always use in order not getting along with full-text, but it should be a pretty straightforward fix.
Thanks for the continuing feedback, all.
ldolse - we really didn't do any testing for this beta with Adso imports (didn't even get Adso converted until a day or two before we released it), so I'm not surprised to see a few problems like this - the other dictionaries are all very meticulous about making sure every character in a headword has a single-character entry, but obviously since Adso's not designed as a printed dictionary and hence don't need to make characters easy to find for radical table lookups it's not as big a concern for them. But yeah, we probably should add a dictionary switch button to that screen.
daniel123 - you're darn right we're not that committed to Palm, frankly the biggest mistake I've made in the history of running Pleco was deciding to put 2.0 on Palm OS; if we'd made it WM-only you'd all be using an uber-stable Pleco 2.1 on your iPAQs and admiring screenshots of the about-to-be-released iPhone version. We're losing large amounts of money for every day we spend working on Palm rather than some other platform with better business prospects. We've made the mistake of telling people for the last few years that they can still go ahead and buy a Palm to run Pleco (this has now been corrected on the PDA recommendations page incidentally), and for the sake of them and our large base of existing users we're doing all we reasonably can to get it working well on Palm, but at some point we have to move on to other projects even if it's not quite as stable and fast as the old version.
So we'll release what we can on Palm, we'll keep offering/supporting 1.0.3 for as long as anybody's interested in it (if there's enough demand we might even consider backporting a couple of features like audio and the new dictionaries to the old system), but if we lose a few customers or a few upgrade purchases because it's less than it could be on Palm, that's a reasonable price to pay for finally getting out of the Development Cycle Which Will Not End and moving on to the many other projects which we really ought to be working on. We made a lot of mistakes in planning 2.0, but even worse than those mistakes would be destroying Pleco's future in an attempt to fix every last one of them - at some point the best strategy is to apologize for what you can't fix, take your lumps and move on.
Re your specific issues, we'll see what we can find with this SlovoEd bug - hopefully it'll prove to be something minor / easy to find (as most such conflicts are). Audio's still a bit unstable after large numbers of plays, as a few other people have mentioned. The bracket-stripping is a bug in the flashcard creation function - it's supposed to remove rounded brackets from Pleco-supplied dictionaries like the ABC et al, since they're always just character variants like (兒), but it's not supposed to do so with user-created dictionary headwords.
The flashcard manager is a bit of a jump from 1.0 but I think once you get the hang of it it's really just as usable. Given the large negative response we've gotten to it, though, we are considering bringing back the old interface from 1.0 too (as an "Organize Flashcards" screen) - there'd still only be one copy of the flashcard in the system (unless you told it to be otherwise) but you could now move / copy a link to it between categories just like in the old version.
Full-text search isn't supported in user dictionaries, unfortunately - just not technically feasible on Palm to generate/update a full-text dictionary index on the fly. (would be on PPC but that'll have to wait for a later version) Not sure why it could be disappearing, might be the file corruption issues we're seeing elsewhere with the TX - hopefully we'll get to the bottom of those soon. (for what it's worth, I essentially haven't left my computer except to sleep in the last 3 days working on almost nothing but this TX bug)
sych - the audio system still isn't quite perfect, yeah - we made it a lot better since Beta 4 but there are still a few occasional memory leaks we're trying to dig up.
westmeadboy - that's actually what it's supposed to do, oddly enough (and what it's done for several versions now) - basically it works that way so that you can use it both to open the fullscreen input dialog and to open/close or change the position of the palette version. (since you might potentially want a larger screen for some inputs but be OK with the palette for others) Since we think most people will now be configuring the input palette using Preferences (much more sensible/intuitive anyway) we thought we might as well leave it working the way it already did in 1.0 for the sake of people who'd gotten used to it. (a logic we probably could have used in a few other places in 2.0 in hindsight)
grabolous - yikes, that's not good; the Messaging data must not have been saved out to the Treo's flash memory correctly when that reset happened. The 650 had a very early version of Palm's flash memory storage system so problems like that aren't as rare as we'd like them to be...
Mator - yeah, we've noticed this in a few places actually - a quirk of Palm OS software design is that the OS sends you through a completely different set of code when you exit a dialog box deliberately (thanks to a button press) than when it forces you to exit, so one of the many little debugging tasks we still haven't gotten around to yet is trying every one of those dialogs with a forced exit and making sure it doesn't crash. Not sure why it's crashing instead of gracefully exiting those no-card sessions but we've gotten reports of that even on PPC so it seems pretty widespread. The static in the woman audio isn't something we can do much about, unfortunately, the original recordings have the same static. Hadn't noticed that bug with always use in order not getting along with full-text, but it should be a pretty straightforward fix.
Thanks for the continuing feedback, all.
Sweet as.
Jesus - that T|X bug must be driving you crazy... I for one am reasonably glad that you supported Palm OS with 2.0, but I'll probably switch to Windows Mobile within the next year or so. Not financially a good idea right now (Smartphone prices in New Zealand are obscene and I'm still a poor student), but with POS basically dead there's no real choice, especially if Pleco 2.1 will be Win-Mob only. Hoping the Treo 850 gets as good a reception as the Treo 800w has.
I am impressed at the stuff you've managed to get Palm OS to do though, however much effort it took.
Hell, even Kinoma and PocketTunes are heading into the WinMob market now...
EDIT: When you drop Palm OS support, will the WM interface be re-designed? I think sfrrr complained a while back that it wasn't really a proper WM interface, but you use the same code for Palm OS and WM, so it can't really be changed.
Jesus - that T|X bug must be driving you crazy... I for one am reasonably glad that you supported Palm OS with 2.0, but I'll probably switch to Windows Mobile within the next year or so. Not financially a good idea right now (Smartphone prices in New Zealand are obscene and I'm still a poor student), but with POS basically dead there's no real choice, especially if Pleco 2.1 will be Win-Mob only. Hoping the Treo 850 gets as good a reception as the Treo 800w has.
I am impressed at the stuff you've managed to get Palm OS to do though, however much effort it took
.Hell, even Kinoma and PocketTunes are heading into the WinMob market now...
EDIT: When you drop Palm OS support, will the WM interface be re-designed? I think sfrrr complained a while back that it wasn't really a proper WM interface, but you use the same code for Palm OS and WM, so it can't really be changed.
Some promising news on the TX front - we just went and bought a brand new TX and it actually looks like the old TX may have been defective; its flash memory seems to have had some faulty cells, so essentially databases were working just fine as long as they stayed in the RAM database cache but once they were written out to flash memory they'd often become corrupted, so that when you retrieved them again there'd be missing records or other problems. (we'd thought this was become we were screwing up the data while it was still in the cache, or preventing it from being written out to disk, but it turns out it was all Palm's fault in this case) What finally got us to consider the possibility of a hardware defect was that we were running into similar sorts of data corruption writing data to the flash memory directly, meaning that either there were two independent TX-specific data corruption bugs (neither of which was affecting any of our other test systems) or there had to be something wrong with the specific TX we were testing on. And since the new TX imports and exports data beautifully without any corruption problems it seems like it's not in fact a TX-wide bug after all.
As I'm sure many TX owners would agree, the TX is not exactly a paragon of hardware quality - buttons breaking, digitizer drifting, etc - and it seems like Palm may have skimped on a bit on quality control for their flash memory as well. We haven't been able to dig up any memory tester utilities for the TX yet but if anybody wants a fun little Palm OS programming project that might be a good one (for other Palms too) - would help with making your case to incredulous Palm warranty replacement folks anyway. Or we could just write one ourselves if we run into a lot of people seeing these sorts of problems - you basically just write out a sequence of numbers to flash memory and then read it back into make sure it hasn't changed, pretty easy.
So lesson learned there, I guess... myself I'm pretty ticked off at having just spent several days pounding away at what turned out to be largely a problem of Palm's cheap-o manufacturing, but it has at least provided the chance to make some other improvements to the Palm OS import and database code - for example, there's now a nice helpful Stop button, so if you suddenly decide you need to do something else with your Palm you can cancel, undo, exit, and try it again later without having to soft reset. It also does a better job of compacting records, reducing file size overhead, it seems to be a bit faster and it supports internal flash memory writes, though whether or not we'll actually expose that functionality we're not sure about yet. (good bit faster for imports at least, but still not as well-tested as we'd like)
Looking through the TX issues reported here, it seems like most of them are probably more ordinary sorts of beta version bugs - daniel123's missing user dictionary database is a bit worrisome, are you sure you'd checked the "store imported defns in user dict" box in the importer screen? If not, those definitions would have been stored in the flashcard database instead (so the user dictionary would be empty) - there's a function in Manage Flashcards you can use to transfer them all to the user dictionary. If anyone runs into a corrupted database / missing data on Palm (and particularly the TX), send us a copy and we'll go in and check to see whether it's the sort of corruption that's likely to have been caused by a faulty memory cell. (usually means it's confined to a specific segment of the file)
ipsi - yeah, we do need to retool the WM interface a bit - the real thing that's going to help with that is doing a desktop version, at that point we'll be making a lot of the UI code less aggressively cross-platform than it is now (WM users may not like having a program that feels too Palm-ish, but Mac users will absolutely clobber you for releasing one that feels too Windows-ish) and hence will be able to do things like putting cutesy little dot-circles around tap-hold selections. We're not going to be able to go back to using standard Microsoft UI controls unless and until they finally get around to fixing the bugs in their font system, but we can at least do a better job of tweaking our own custom controls to be more Windows-like. (I think in general 2.0 makes a lot of progress on that front, though)
As I'm sure many TX owners would agree, the TX is not exactly a paragon of hardware quality - buttons breaking, digitizer drifting, etc - and it seems like Palm may have skimped on a bit on quality control for their flash memory as well. We haven't been able to dig up any memory tester utilities for the TX yet but if anybody wants a fun little Palm OS programming project that might be a good one (for other Palms too) - would help with making your case to incredulous Palm warranty replacement folks anyway. Or we could just write one ourselves if we run into a lot of people seeing these sorts of problems - you basically just write out a sequence of numbers to flash memory and then read it back into make sure it hasn't changed, pretty easy.
So lesson learned there, I guess... myself I'm pretty ticked off at having just spent several days pounding away at what turned out to be largely a problem of Palm's cheap-o manufacturing, but it has at least provided the chance to make some other improvements to the Palm OS import and database code - for example, there's now a nice helpful Stop button, so if you suddenly decide you need to do something else with your Palm you can cancel, undo, exit, and try it again later without having to soft reset. It also does a better job of compacting records, reducing file size overhead, it seems to be a bit faster and it supports internal flash memory writes, though whether or not we'll actually expose that functionality we're not sure about yet. (good bit faster for imports at least, but still not as well-tested as we'd like)
Looking through the TX issues reported here, it seems like most of them are probably more ordinary sorts of beta version bugs - daniel123's missing user dictionary database is a bit worrisome, are you sure you'd checked the "store imported defns in user dict" box in the importer screen? If not, those definitions would have been stored in the flashcard database instead (so the user dictionary would be empty) - there's a function in Manage Flashcards you can use to transfer them all to the user dictionary. If anyone runs into a corrupted database / missing data on Palm (and particularly the TX), send us a copy and we'll go in and check to see whether it's the sort of corruption that's likely to have been caused by a faulty memory cell. (usually means it's confined to a specific segment of the file)
ipsi - yeah, we do need to retool the WM interface a bit - the real thing that's going to help with that is doing a desktop version, at that point we'll be making a lot of the UI code less aggressively cross-platform than it is now (WM users may not like having a program that feels too Palm-ish, but Mac users will absolutely clobber you for releasing one that feels too Windows-ish) and hence will be able to do things like putting cutesy little dot-circles around tap-hold selections. We're not going to be able to go back to using standard Microsoft UI controls unless and until they finally get around to fixing the bugs in their font system, but we can at least do a better job of tweaking our own custom controls to be more Windows-like. (I think in general 2.0 makes a lot of progress on that front, though)
hleperlier
举人
I was travelling from Taiwan to Europe when Beta 4 expired. Mike, anyway to export flashcards I have created with Beta 4 (as xml file for instance) and then import them into beta 5.
I have downloaded Beta 5 and will install it today on my Palm TX. A bit worried that Palm TX seems causing so many problems
thanks
I have downloaded Beta 5 and will install it today on my Palm TX. A bit worried that Palm TX seems causing so many problems
thanks
hleperlier
举人
Sorry for the bad spelling and writing: I meant: is there any way to export flashcards I have created with Beta 4 (as xml file for instance) and then import them into beta 5?
Thanks
Thanks
Some flashcard issues
Hi there. this sure is an excellent program! I really love the flashcards and their complexity! That was just what I had been looking for.
Yes, it did mess up my flashcards, but I don't want to try to reproduce that.
However, some "features" I am able to reproduce:
Oh, and is there a way to transfer scores from one score file to another? I have scorefiles for reading, writing, speaking and listening. All my previous scores (from 1.0) were copied into reading, but for some, I'd like to have them in listening, for example.
Thanks again for the excellent programming work!
Hi there. this sure is an excellent program! I really love the flashcards and their complexity! That was just what I had been looking for.
Yes, it did mess up my flashcards, but I don't want to try to reproduce that.
However, some "features" I am able to reproduce:
- I am not able to untick the "only change scores once per day" button.
- importing worked ok for most flashcards, but for some, one or two digits were added to the (repetition spacing based) score. not too much trouble to divide those scores by 10 or 100 respectively, though.
- in "manage flashcards" I have not been able to filter for two categories at the same time (AND). Want to do this because I want to apply some operations to all flashcards that were in a certain "list" and had a certain "flag" (both of which are now categories). [Eventually I did it with SQLite - thanks for the straightforward DB-Design - but I guess that's not the proper way to do it...]
Oh, and is there a way to transfer scores from one score file to another? I have scorefiles for reading, writing, speaking and listening. All my previous scores (from 1.0) were copied into reading, but for some, I'd like to have them in listening, for example.
Thanks again for the excellent programming work!
>daniel123's missing user dictionary database is a bit worrisome, are you sure you'd checked the "store imported defns in user dict" box in the importer screen?
Well, one of the soft resets leads to one of the known neverending Palm TX reboots where I just can do a hard reset to get out. For I didn't already make a hotsync after flashcard import I lost the flashcards and had to import it again (no problem). Anyway this time the user dictionary works fine. Maybe I did a mistake the last time. So we can forget about it.
I found another problem with the scetch box for flashcard learning: The handwriting doesn't seem to display correctly. It looks like that it looses something. Another but not important problem with xml export from exporter: In the headwort and the pronunciation I somtimes use real numbers because I have often complete sentences I want to learn that contain for example date or time. The numbers in the pronunciation seem not to be exported but ignored.
To say a few words about Windows Mobile vs Palm. When I started with Pleco I could read everywhere in the Internet that the Palm version is much more stable. That's why I decided for the Palm. Some months ago I bought a TX just for beta testing and to use it later for the new release. If I had known that Windows Mobile would be the better choice it would be no problem but now it's a bit annoying.
But the TX is not a good choice also because of other reason: after three or two soft reboots it usually leads to the neverending reboot where I had to do a hard reset to get out and always I lose all data that change after last sync. This means on a TX it's only possible to use very stable software. So it couldn't be seriously used for a software that soft reboots from time to time. Even if PD 2.0 would be a little more stable it is still dangerous to use it in real life unless it's as stable as today. Of course it's not PD's problem.
At the moment the beta5 soft resets about after 3 to 8 Minutes using. It seems that it doesn't matter what I do (using flashcard or play with the dicts).
So maybe I will consider to try PD 2.0 on a iPaq in future if this would be really better. Most problem is to move and preserve all existing data in calendar, contacts and memos. I also found some comments that the ipaq often has syncing problems. This doesn't sound nice. I am afraid to come from one problem to another. And the question for me is also if it works also good under Vista. Well best choice would be to wait a time and look how PD envolves in future before I buy a new device.
Daniel
Well, one of the soft resets leads to one of the known neverending Palm TX reboots where I just can do a hard reset to get out. For I didn't already make a hotsync after flashcard import I lost the flashcards and had to import it again (no problem). Anyway this time the user dictionary works fine. Maybe I did a mistake the last time. So we can forget about it.
I found another problem with the scetch box for flashcard learning: The handwriting doesn't seem to display correctly. It looks like that it looses something. Another but not important problem with xml export from exporter: In the headwort and the pronunciation I somtimes use real numbers because I have often complete sentences I want to learn that contain for example date or time. The numbers in the pronunciation seem not to be exported but ignored.
To say a few words about Windows Mobile vs Palm. When I started with Pleco I could read everywhere in the Internet that the Palm version is much more stable. That's why I decided for the Palm. Some months ago I bought a TX just for beta testing and to use it later for the new release. If I had known that Windows Mobile would be the better choice it would be no problem but now it's a bit annoying.
But the TX is not a good choice also because of other reason: after three or two soft reboots it usually leads to the neverending reboot where I had to do a hard reset to get out and always I lose all data that change after last sync. This means on a TX it's only possible to use very stable software. So it couldn't be seriously used for a software that soft reboots from time to time. Even if PD 2.0 would be a little more stable it is still dangerous to use it in real life unless it's as stable as today. Of course it's not PD's problem.
At the moment the beta5 soft resets about after 3 to 8 Minutes using. It seems that it doesn't matter what I do (using flashcard or play with the dicts).
So maybe I will consider to try PD 2.0 on a iPaq in future if this would be really better. Most problem is to move and preserve all existing data in calendar, contacts and memos. I also found some comments that the ipaq often has syncing problems. This doesn't sound nice. I am afraid to come from one problem to another. And the question for me is also if it works also good under Vista. Well best choice would be to wait a time and look how PD envolves in future before I buy a new device.
Daniel
Mike,
Sorry for harping on this, I'm just not sure I'm getting my point across. I'm not trying to point out that you should fix Adso or anything like that. My point is that the current import algorithm seems rather lazy - if 'something' matches in the headwords/pinyin, even if it matched ambiguously, then import routine assumes that that is a usable match. If you have absolute control over every single entry of every single dictionary entry then that's ok. The reality is that even the pay dictionaries could have a problem that was missed during the import process, and the free dictionary problem will only get worse as you add CEDICT, LDS, Handedict, etc. You shouldn't have to do editorial work on the free dictionaries, but mistakes in those dictionaries shouldn't be messing up flashcard imports. The other reality is that the free dictionaries actually have some of the best entries that lend themselves to flashcard review, particularly multiple choice. That's the reason I plan to use them as my primary flashcard import dictionaries.
My thinking is that the matching algorithm shouldn't be lazy, it should be exact. If the flashcard file headword is a single character then the matching algorithm should only be allowing dictionary headwords with a single character. Non-chinese characters in the headword should count, like the Daihatsu entry that had ’大？‘ － that's a two character headword. If the flashcard file has both a Hanzi headword and Pinyin then both the Pinyin syllables AND the headword should match. The only situation where ambiguity seems acceptable to me is when the pinyin tones don't match.
I don't think another button to go to the next dict is required if the ambiguity issue was resolved. Eliminating ambiguous matches would reduce the number of times user intervention is required during the import, and frankly I'd like to reduce that whereever possible. Right now importing any of the flashcard lists is a lot of work when Adso is the first dictionary in the list. Hopefully CEDICT will reduce the number of headword duplicates, I realize that's the nature of how Adso works.
Sorry for harping on this, I'm just not sure I'm getting my point across. I'm not trying to point out that you should fix Adso or anything like that. My point is that the current import algorithm seems rather lazy - if 'something' matches in the headwords/pinyin, even if it matched ambiguously, then import routine assumes that that is a usable match. If you have absolute control over every single entry of every single dictionary entry then that's ok. The reality is that even the pay dictionaries could have a problem that was missed during the import process, and the free dictionary problem will only get worse as you add CEDICT, LDS, Handedict, etc. You shouldn't have to do editorial work on the free dictionaries, but mistakes in those dictionaries shouldn't be messing up flashcard imports. The other reality is that the free dictionaries actually have some of the best entries that lend themselves to flashcard review, particularly multiple choice. That's the reason I plan to use them as my primary flashcard import dictionaries.
My thinking is that the matching algorithm shouldn't be lazy, it should be exact. If the flashcard file headword is a single character then the matching algorithm should only be allowing dictionary headwords with a single character. Non-chinese characters in the headword should count, like the Daihatsu entry that had ’大？‘ － that's a two character headword. If the flashcard file has both a Hanzi headword and Pinyin then both the Pinyin syllables AND the headword should match. The only situation where ambiguity seems acceptable to me is when the pinyin tones don't match.
I don't think another button to go to the next dict is required if the ambiguity issue was resolved. Eliminating ambiguous matches would reduce the number of times user intervention is required during the import, and frankly I'd like to reduce that whereever possible. Right now importing any of the flashcard lists is a lot of work when Adso is the first dictionary in the list. Hopefully CEDICT will reduce the number of headword duplicates, I realize that's the nature of how Adso works.
hleperlier - sorry about the unexpected expiration, we should probably put a note about that into the download file (or display it on first run) but honestly we thought 2.0 would be done by 8/1. B5 expires on 10/1 for everyone's info. The number of TX problems reported here is I think largely a function of how many people are using TXes - for a "power user" who wants a Palm but doesn't want a smartphone they're pretty much the only option, but having resolved that file corruption issue I don't really think the situation on them is that much worse than on other Palms.
The XML format has changed a bit between B4 and B5, but as long as you're not expecting 100% accuracy, an XML export from B4 should come into B5 reasonably smoothly - you might end up with a few weirdly-named categories but it should be mostly OK aside from that.
marsch - how did it mess up your flashcards? Reproducible or no, file corruption is definitely something we want to know about. Not sure why the once-per-day switch isn't working, we made that optional rather recently (used to be automatic but then we realized some people might not like it). We've seen that extra digits issue one or two other times, actually - probably a lazily-cleared text buffer or something along those lines. You can do an AND filter in Manage Flashcards by going to the Settings screen. (but yeah, the SQL design is pretty straightforward - we'll probably even release some basic documentation on it at some point, at least for the category/card/score areas which people are likely to want to hack away on) I assume from the fact that you were able to edit the SQLite database that you're on Windows Mobile, right?
Send me an e-mail with the new SD card ID and I'll switch your license over to that. (same goes for anybody else who wants to try that out) There's no way to transfer individual scores between files, but if you duplicate a scorefile that would give you a second copy of all of those scores that you could then use for other things - if you don't want scores for cards you're not interested in taking up space, make sure that scorefile's the active one, pull up all of the cards you're not interested in in a Manage result and then reset their statistics. (or just go into the SQLite file and delete any record in that scorefile's table with a card ID that you're not interested in - cards only get score records when they've been reviewed at least once, so it's pretty much always safe to delete score records)
daniel123 - yeah, we've seen quite a few of those neverending loops here too. Resets in general seem significantly less common after the changes we've made in the last few days, though, even on the defective TX... glad to hear the user dictionary's working correctly this time at least. Not sure why the sketch box isn't working correctly - do you see the same behavior with handwriting input or only with sketches? The numbers in pronunciation are getting stripped out because it doesn't want them to be confused with tone numbers - I guess the best solution on that might be to offer the option to export Pinyin with tone marks instead, or just to apply the don't-try-to-clean-up-user-created-entries rule to pronunciation as well as headwords.
The Palm version of Pleco 1.0 is in fact more stable, and to me at least also feels a good bit faster - the biggest reason why we've still been recommending Palm the last few years is because for 1.0 it really is the best platform to run Pleco on. There was also that few-month window in which we had that nasty Windows Mobile 6 crashing bug we couldn't figure out how to fix, steering even more people towards Palm. If we'd done a better job the first time around with Windows Mobile we might have been recommending that instead, but as is, for anyone who wanted to use Pleco before now Palm really was the better bet. Outside of flashcards I still like Palm better than WM in some respects - the resolution on Palm screens (especially Treos) is just right for displaying characters clearly without anti-aliasing, the brighter colors go great with our new headword-coloring feature, Instant Access though slower to start up is noticeably easier to access on Palm than on WM, and as others have pointed out Pleco's interface is still a bit more Palm-like than Windows-like. So I don't feel too bad that we've been recommending Palm all this time, since it legitimately has been the better platform to run Pleco on - we really just should have done a better job of making it clear that that wouldn't necessarily remain the case with Pleco 2.0. But I'm sorry you ended up buying a TX for the sake of 2.0.
The best way to preserve calendar/contact/etc data going between Palm and WM is to sync your Palm with Outlook (if you don't have Outlook, it's included with almost every Pocket PC and almost every Palm includes software to get Palm Desktop syncing with it) - Vista shouldn't be much of a problem. Haven't heard much about these sync issues on iPAQ, in general I like Palm's sync-when-I-tell-you-to model over WM's sync-when-you-feel-like-it but we haven't noticed any specific issues with sync from our iPAQs. (though the 110's scroll down button just recently stopped working, so maybe HP has some of the same quality control problems that Palm does...)
ldolse - oh I do understand that we need to get it to disambiguate better, but it's still largely a free-dictionary-specific issue. The importer actually doesn't assume that any partial match is correct, it just assumes that with a single-character search the first result it gets will always be a single-character entry, something we can guarantee with our own dictionaries but not with Adso. The reason it accepts inaccurate Pinyin matches is because in a lot of cases there's significant disagreement between one dictionary and another about Pinyin, and not even just tones; a lot of flashcard lists also tend to be less-then-perfectly proofread, so we frequently run into Pinyin errors in those. The current algorithm basically checks every entry it finds for both partial and exact matches; if it finds an exact match it uses that (or an ambig prompt if it finds more than one), if it only finds a partial match (i.e. headword is OK but pinyin or traditional characters don't line up) it uses that (or again an ambig prompt if it finds more than one), if it doesn't find any matches it moves on to the next dicitonary.
Rejecting outright any result that doesn't match the length correctly makes sense (though we do need to ignore punctuation marks like ?, because quite a few dictionaries include those in headwords and we don't want them factored into the length in that case), and we should probably also check other dictionaries for exact matches when we only find partial ones from a particular dictionary, but these are both mainly important for people using something other than the default import dictionary order. If you're not using Adso and ABC is your first dictionary then pretty much everything is going to match the ABC and you'll end up with a quite acceptably accurate set of flashcards. So this is why we didn't put too much stress into those two areas for the beta - in hindsight we probably should have waited for the finished version to release Adso at all, since at that point we'll have actually done some testing with it along with the other dictionaries.
The XML format has changed a bit between B4 and B5, but as long as you're not expecting 100% accuracy, an XML export from B4 should come into B5 reasonably smoothly - you might end up with a few weirdly-named categories but it should be mostly OK aside from that.
marsch - how did it mess up your flashcards? Reproducible or no, file corruption is definitely something we want to know about. Not sure why the once-per-day switch isn't working, we made that optional rather recently (used to be automatic but then we realized some people might not like it). We've seen that extra digits issue one or two other times, actually - probably a lazily-cleared text buffer or something along those lines. You can do an AND filter in Manage Flashcards by going to the Settings screen. (but yeah, the SQL design is pretty straightforward - we'll probably even release some basic documentation on it at some point, at least for the category/card/score areas which people are likely to want to hack away on) I assume from the fact that you were able to edit the SQLite database that you're on Windows Mobile, right?
Send me an e-mail with the new SD card ID and I'll switch your license over to that. (same goes for anybody else who wants to try that out) There's no way to transfer individual scores between files, but if you duplicate a scorefile that would give you a second copy of all of those scores that you could then use for other things - if you don't want scores for cards you're not interested in taking up space, make sure that scorefile's the active one, pull up all of the cards you're not interested in in a Manage result and then reset their statistics. (or just go into the SQLite file and delete any record in that scorefile's table with a card ID that you're not interested in - cards only get score records when they've been reviewed at least once, so it's pretty much always safe to delete score records)
daniel123 - yeah, we've seen quite a few of those neverending loops here too. Resets in general seem significantly less common after the changes we've made in the last few days, though, even on the defective TX... glad to hear the user dictionary's working correctly this time at least. Not sure why the sketch box isn't working correctly - do you see the same behavior with handwriting input or only with sketches? The numbers in pronunciation are getting stripped out because it doesn't want them to be confused with tone numbers - I guess the best solution on that might be to offer the option to export Pinyin with tone marks instead, or just to apply the don't-try-to-clean-up-user-created-entries rule to pronunciation as well as headwords.
The Palm version of Pleco 1.0 is in fact more stable, and to me at least also feels a good bit faster - the biggest reason why we've still been recommending Palm the last few years is because for 1.0 it really is the best platform to run Pleco on. There was also that few-month window in which we had that nasty Windows Mobile 6 crashing bug we couldn't figure out how to fix, steering even more people towards Palm. If we'd done a better job the first time around with Windows Mobile we might have been recommending that instead, but as is, for anyone who wanted to use Pleco before now Palm really was the better bet. Outside of flashcards I still like Palm better than WM in some respects - the resolution on Palm screens (especially Treos) is just right for displaying characters clearly without anti-aliasing, the brighter colors go great with our new headword-coloring feature, Instant Access though slower to start up is noticeably easier to access on Palm than on WM, and as others have pointed out Pleco's interface is still a bit more Palm-like than Windows-like. So I don't feel too bad that we've been recommending Palm all this time, since it legitimately has been the better platform to run Pleco on - we really just should have done a better job of making it clear that that wouldn't necessarily remain the case with Pleco 2.0. But I'm sorry you ended up buying a TX for the sake of 2.0.
The best way to preserve calendar/contact/etc data going between Palm and WM is to sync your Palm with Outlook (if you don't have Outlook, it's included with almost every Pocket PC and almost every Palm includes software to get Palm Desktop syncing with it) - Vista shouldn't be much of a problem. Haven't heard much about these sync issues on iPAQ, in general I like Palm's sync-when-I-tell-you-to model over WM's sync-when-you-feel-like-it but we haven't noticed any specific issues with sync from our iPAQs. (though the 110's scroll down button just recently stopped working, so maybe HP has some of the same quality control problems that Palm does...)
ldolse - oh I do understand that we need to get it to disambiguate better, but it's still largely a free-dictionary-specific issue. The importer actually doesn't assume that any partial match is correct, it just assumes that with a single-character search the first result it gets will always be a single-character entry, something we can guarantee with our own dictionaries but not with Adso. The reason it accepts inaccurate Pinyin matches is because in a lot of cases there's significant disagreement between one dictionary and another about Pinyin, and not even just tones; a lot of flashcard lists also tend to be less-then-perfectly proofread, so we frequently run into Pinyin errors in those. The current algorithm basically checks every entry it finds for both partial and exact matches; if it finds an exact match it uses that (or an ambig prompt if it finds more than one), if it only finds a partial match (i.e. headword is OK but pinyin or traditional characters don't line up) it uses that (or again an ambig prompt if it finds more than one), if it doesn't find any matches it moves on to the next dicitonary.
Rejecting outright any result that doesn't match the length correctly makes sense (though we do need to ignore punctuation marks like ?, because quite a few dictionaries include those in headwords and we don't want them factored into the length in that case), and we should probably also check other dictionaries for exact matches when we only find partial ones from a particular dictionary, but these are both mainly important for people using something other than the default import dictionary order. If you're not using Adso and ABC is your first dictionary then pretty much everything is going to match the ABC and you'll end up with a quite acceptably accurate set of flashcards. So this is why we didn't put too much stress into those two areas for the beta - in hindsight we probably should have waited for the finished version to release Adso at all, since at that point we'll have actually done some testing with it along with the other dictionaries.