The PlecoDict Beta Feedback Thread

Minor quirks

First, the beta looks very polished and my overall experience has been very good.

Some minor things I've discovered:

* search for bai3 in the Oxford C-E and you will see an entry starting with a trademark sign and a three-four quarters sign. Probably something went south during convertion of the dictionary.

* xiang3 (noisy) in the ABC: the roman III has a problem: it consists of a bold II and a non-bold final I. Nothing the next convertion couldn't solve.
 
A

Anonymous

Guest
Flashcards

Hi Mike

I still haven't properly acquainted myself with the flashcard module options, but while you're still tweaking them, could I suggest the addition of a timed auto-display mode that cycles through cards (with/without subsequent definitions) - enabling hands-free operation. I notice that you have included a timer, so perhaps this might be feasible.
I appreciate what you have said about the necessary size of the icons and I have found that after eliminating the spaces, I could configure them exactly as I had wanted. Also, my elusive command bar icon has now put in an appearance.

Once again congratulations on a fine job!

Baoluo
 
A

Anonymous

Guest
..cont'd

...I would also like to suggest the centering of the characters within the display area and is it possible to scale them up to an even bigger and bolder size to dominate the available (or entire :D ) screen area?

Regards
Baoluo
 
flashcard import

The import of UTF-8 text files that contains certain european accented characters doesn't seem to work properly. For instance, german ?, ?, ?, et cetera, end up as something different (two different characters) after import.

I tried to import an small text file from SD card containing ?, et cetera, and the u-umlaut didn't made it through.
 
damaged user dictionary C-E

I first imported a few flashcards from SD card. This imports the flashcards into the user dictionary at the same time (so to say).

Then I switched to the user dictionary and switched between simplified and traditional chinese character sets.

After switching back and forth the list showed garbled entries with the same chinese character being displayed all along the list. Leaving and restarting PlecoDict did not solve the problem, so I had to delete the user dictionary database.

BTW - is it possible to move the user dictionary to SD card?
 

gandq

探花
G-C encoding


once had the same thing happening that harald discribes above, but i don't remember if it happened after i switched character sets or somehow else.


the encoding issue that harald mentions is very tricky. i think there just are conflicts between german and chinese encoding that exist even on my notebook (taiwanes win XP).

still, plecodict can show german umlaute and chinese characters at the same time. the umlaute will be replaced with some gibberish or ?? after i import a file (or maybe this already happens on the notebook while converting text to a pdb - will check this), but i can then manually enter the german chars in plecodict and have them displayed correctly along with the chinese chars. interesting thing is: if i go to the 'edit custom entry' menu again, they will be messed up again or even have just disappeared. strange.

but as said, encoding is a very tricky thing, especially in the combination german + chinese.

i don't really unterstand how all that encoding works. do you have any suggestions on this, mike?


thanks,



jo.
 

curwenx

秀才
curwenx - thanks for the crash report, we'll investigate this and fix it for 1.0. The enable search button is actually doing what it's designed to, but we should probably add a "button only" option for search delay - the current system is a bit silly. And with flashcard import, did you create the wordlist on your PC or on your Palm? And are you sure you configured the text encodings correctly?

My bad on the word list. It was a text file created on my PC, but I saved it as regular Unicode, and found out later that Plecodict expects UTF-8. Once I fixed it, I was able to import. The file contained 500 single characters, and PlecoDict crashed on a NULL pointer after working for a while (and was unable to import all the entries).

Also, if I enter an incorrect filename, pressing the import button does nothing, I guess there's should be a "file not found" warning. Also, notes I have in MemoPad do not show up in the import dialog. Furthermore, when importing, no status feedback is given.

I'd also like to mention another unrelated problem I've been having. Whenever I try to bring up regular keyboard input by tapping in the lower left corner I have to tap multiple times (random) leaving a trail of spaces before the soft keyboard eventually shows up. I guess a toolbar button for bringing up the soft keyboard would be nice to have as well..

Also, Auto-detect Copy to Input lang. doesn't seem to work at all. If I select a Chinese word while in the English-Chinese dictionary and copy it to the input area I do not get the proper Chinese word definition, I have to manually press the zhong button first.
 

gandq

探花
yep, there should definitely be some status feedback while importing.

wanted to mention this, too.

with regard to the encoding, i haven't succeeded in importing traditional character lists yet. but again, this could be either be related to my desktop settings or to the chinese OS settings. will do further research on that.


greetings,




jo.
 

chaiusa

举人
Mike,

now i got T5.
I try to put the database files in INTERNAL /applications/ folder. but it seem like plecodict can't read from this location.

Please help
 

mikelove

皇帝
Staff member
haraldalbrecht - with the bai3 bug, I'm not seeing that here, are you sure you're using the latest version? And is it definitely bai3? The xiang3 bug is actually in the Oxford C-E dictionary, but thanks for pointing it out.

I'm surprised to hear you're having such serious problems with the user dictionary, I thought it was relatively stable at this point - we'll keep working on it. We never even tested the UTF-8 importer on European text, so it's very likely there's some sort of bug there.

I'm afraid there's no way to put the user dictionary on an SD card; the Palm OS functions for reading/writing from SD cards are dreadfully awkward and even if we did get it working it would be so slow as to be almost useless. However, it should be possible (at least in the finished version) to copy one user dictionary to your SD card and start a brand new one (without breaking any flashcard links), so if it starts eating up too much of your storage space that would be worth considering.

gandq - the characters are probably getting messed up because PlecoDict is trying (and failing) to treat them as Pinyin syllables - the converter for that is a bit crude and as I said before we haven't done a lot of testing with non-English character sets yet. But we certainly will be... about all I can suggest now is that you use the custom dictionary option sparingly until we get this stuff fixed.

curwenx, the null pointer problem might be related to a lack of memory - how much free memory is there on your Palm? The lack of a "file not found" message is something we'll definitely be fixing. A keyboard toolbar button is a good idea, though I can't think of a reason why our software would make it unusually difficult to bring up the soft keyboard in the usual way... we'll look into it, though.

And did you enable "input field compatibility mode"? The auto-detect feature actually doesn't work with that at the moment - it may be possible to fix this in the finished verison, but at the very least we'll disable it by default when input field compatibility is enabled.

chaiusa - are you sure you're using the newest version of PlecoDict? Applications is one of the three directories it normally searches by default. Try moving them to /Palm/Programs/PlecoDict and see if that helps, you can do that easily if you put your T5 in Drive Mode.
 

curwenx

秀才
curwenx, the null pointer problem might be related to a lack of memory - how much free memory is there on your Palm? The lack of a "file not found" message is something we'll definitely be fixing. A keyboard toolbar button is a good idea, though I can't think of a reason why our software would make it unusually difficult to bring up the soft keyboard in the usual way... we'll look into it, though.

My Palm is kind of low on memory, yes, it has about 1 meg free. I thought the soft keyboard "bug" was kind of strange too, and I've noticed that it actually shows up in other apps as well (took me some time to notice since I now use my palm almost exclusively as a dictionary).
Maybe it's related to CJKOS? Of course, having a toolbar button will make me forget all about it :)

And did you enable "input field compatibility mode"? The auto-detect feature actually doesn't work with that at the moment - it may be possible to fix this in the finished verison, but at the very least we'll disable it by default when input field compatibility is enabled.

Yes, I'm using input field compatibility mode, since I use PenPower to write characters. When you eventually incorporate your own handwriting recognizer, I'll probably switch to that provided it provides the same freedom for cursive writing as penpower does.

I'm looking forward to the release of the dictionary creation tool. Will this allow me to add both C-E and E-C dictionaries?
 

Henry

进士
"springing-back bug"

And do you currently have the search delay feature turned on? That seems like the most likely culprit for this springing-back bug, it's about the only part of the software where something happens after a 1-second delay.

It does seem to be related to the search delay feature-

When I have the delay feature turned on, it 'springs-back' after the number of seconds I have selected to delay.

If I turn off the delay feature, it springs back immediately.

Is this coming up on any other devices or just my T|W?


Other- I'm also having the some of the problems that were mentioned above, namely

Flashcards- under test settings, selecting Test: Char/PY/Defn is not reliable, I cannot test just for Defn.

Soft Reset while in full screen entry list mode and I delete 2nd of 2 characters in entry field.

Some funniness when searching in English, for example, searching for 'window' comes up with 'wind' . . .

Roman Numerals l and lll in the C-E Oxford dictionary sometimes appear bold, sometimes normal, sometimes a combination of the two . . . and the bold roman numeral l often appears too close to the pin-yin entry, appearing to be the letter 'L' (sorry I cannot find an example right now)


One unexpected benefit, I find I no longer have to select the input field (to set the cursor blinking) while writing with Penpower!
 
Re: G-C encoding

gandq said:
the encoding issue that harald mentions is very tricky. i think there just are conflicts between german and chinese encoding that exist even on my notebook (taiwanes win XP).

The encoding problem you are refering to is when mixing ISO8859-1 and GB or BIG5 encodings. However, Plecodict fortunately also allows also for UTF-8 encoding ... and thus avoids mixing up and trashing ISO8859-1 umlauts due to GB/BIG5 encodings also using similiar character codes for encoding plane switching. For that reason I mentioned in my original post that I'm using UTF-8 for import.

BTW: gandq, do you have a stuck bold key?
 
mikelove said:
haraldalbrecht - with the bai3 bug, I'm not seeing that here, are you sure you're using the latest version?

Mike, please excuse me for not being precise enough: it's bai3guo3 that triggers the problematic OX dictionary entry. I removed all files and dictionaries before installing the newest beta since you had already said in advance that the new beta would need to introduce a slightly changed dictionary database format.

mikelove said:
I'm surprised to hear you're having such serious problems with the user dictionary, I thought it was relatively stable at this point - we'll keep working on it. We never even tested the UTF-8 importer on European text, so it's very likely there's some sort of bug there.

I already thought so, that's the reason why I'm trying to give feedback as soon as possible.

mikelove said:
I'm afraid there's no way to put the user dictionary on an SD card; the Palm OS functions for reading/writing from SD cards are dreadfully awkward and even if we did get it working it would be so slow as to be almost useless. However, it should be possible (at least in the finished version) to copy one user dictionary to your SD card and start a brand new one (without breaking any flashcard links), so if it starts eating up too much of your storage space that would be worth considering.

That would help me. I still have a T|T with several applications that unfortunately need to have their data in RAM, so I'm rather tight when it comes to free memory. The T5 has no universal connector which I need for my keyboard. And with its problems here and there the T5 is surely not that interesting...

Will the final version also support flashcards for custom dictionaries? If a converter will be available this could also be a solution: then I can just create personal dictionaries (in German) and use them both for lookup as well as learning.

I saw in the documentation that sorting is not supported for custom dictionaries. Is this going to be resolved, or does this only apply to the user dictionary but not custom dictionaries, which have been created on an external PC?
 

mai

秀才
Hi Mike,

I'm using the Beta on a german T3. Problem is that the instant access feature is not working. When I highlight some text in another app and draw the command stroke there is no Pleco symbol. It worked though in the previous release. CJKOS is activated. Any ideas?
 

mai

秀才
Okay, found it myself (and the message on page 1of this thread telling the same story :). Instant Access is working. There is just no symbol coming up, just a blank space.
 
A

Anonymous

Guest
That 'Missing Icon' issue is a weird one - I notice sometimes that when you tap the space you can JUST make out the icon as it flashes up very briefly.
 

dom

秀才
Expanded CEDICT?

mikelove said:
Actually we've found an expanded version of CEDICT (120,000 entries or so)

Out of curiosity, is this the one hosted at mandarintools.com? Or is it a different one? (How many versions of CEDICT are there?)
 

gandq

探花

@ haraldbrecht: there is no bold key that could get stuck ;) i just like it that way.



@ mike,

interestingly, i was playing around with the user dictionary today (obviously that's what i am most excited about not being an english native speaker) when i found a bug that has to do with bold text:

it seems that sometimes the pinyin of a word won't be shown in bold text if it spans over two rows (varies if the textsize is changed). i can't really describe it. what i did is this:

- i entered 一概而论 and 臭小河马 (as said i was just fooling around), each with pinyin and a definition
- if the textsize is big, the last pinyin syllable 'lun' of the first entry won't be in bold script like the other three syllables (i entered a space between each of them)
- if the textsize is small, the last syllable 'ma' won't be in bold script in the other entry
- both entries will have full bold pinyin if the textsize is changed, respectively

i have other entries where the pinyin spans over more than two rows an where this doesn't happen. strange.

i only encounterd this in the user dictionary.


greetings,




jo.
 

mikelove

皇帝
Staff member
curwenx - this could definitely be CJKOS-related. As far as our own recognizer, though, we're actually not planning to support cursive at all; it's very very difficult to get cursive recognition working correctly, and there are already a lot of other products out there that do it well. Our focus is more on supporting people for whom off-the-shelf Chinese handwriting recognizers might not work so well; people with less-than-perfect stroke order, for example. Some of our designs are really quite different from anything else out there, both in capabilities and interface. For your purposes, I'm afraid you're going to be stuck with PenPower for a while, but we'll do our best to improve support for it (in fact it looks like there's a way we can get newer versions of PenPower working even with Input Field Compatibility disabled).

Henry - we haven't gotten any other reports of this scroll-snapping bug, but hopefully if we try it out on enough Palm OS models we'll find some other one that's afflicted by it. We're aware of those problems in the Oxford data files (the "window" one is actually data-file-related too); we put more effort into cleaning up the NWP/ABC ones for this release, because people who purchased those don't have the option of going back to the older-version software and because we're working on some other big improvements in those files (like the addition of the new 3rd Edition entries), but they should be in much better shape soon. I'm glad to hear that PenPower improvement worked, in general the software should be a lot better about sticking to the Input Field now.

haraldalbrecht - thanks for the clarification, I definitely do see this with bai3guo3. It did seem odd, the newer version shouldn't even be able ot read the older data files (but should simply ignore them). And you can already create flashcards based on user-created dictionaries (just add them to the flashcard list like you would with built-in ones), but it'll certainly be possible to do it with converted ones as well at whatever point we release the converter. We're probably not going to be able to get user dictionaries sorted like the built-in ones (for various reasons, this would be unacceptably slow), but we will at least display them in index order (like we did with the other dictionaries prior to this release) instead of the order in which you added them. Custom dictionaries, however, will be sorted however you want them to be.

mai - yeah, this is a missing icon problem, it actually takes a surprisingly large amount of code to get that little icon to display (Palm OS only expects the currently-running program to put icons in the command bar).

dom - there are actually two expanded versions out there (that we know of), one at http://adsotrans.com/ whose creator is graciously permitting us to use it (as long as we don't charge for it), and another one from UPenn at http://www.ldc.upenn.edu/Projects/Chinese/ (see the wordlists in the "resources" section). I believe the adsotrans one is actually an expanded version of the UPenn one. We'll probably also offer the original CEDICT for people who wish to conserve storage space.

gandq - this is a pretty common place for bugs to appear, actually; we wrote our very own version of the Palm OS text field system (with support for rich text and Unicode) and for performance reasons we store the formatting data for the start of each line in the same data structure that we store word-wrapping calculations in. So it looks like for some reason the bold text formatting isn't carrying over to the next line. Anyway, this should be pretty easy to reproduce and fix.
 
Top