The PlecoDict Beta Feedback Thread

[Anonymous]

a few bugs in the flash cards

hey Mike great job on the Plecodict. I have been anxiously waiting to try out the new version and especially the flash card portion.

Here are some bugs that I have encountered from my use. You had warned that the flash card portion was still buggy and so you may be already aware of these, but I thought I would submit them just incase:

Environment:
T/E with 8.93m of free memory internally,
ABC and NWP loaded on external card with 194.25m free mem

General Bug:
inside Plecodict I do not get any of the instanct access icons (maybe the wrong name for this - little icons at the bottom of the screen that pop up after an upward (L to R) slash) for things like cut , copy, and paste, but I can still do these functions through drop down menus.

Flash Card Bugs:
-On the New Session screen, the Manage Lists button does nothing

-When setting up a flash card session, and if I try to use Freq Adjust with Fixed # of cards, the palm will freeze, and need to be rebooted

-Inport problems:
- a character with multiple pronunciations would sometimes change. For example I added 读 d? to my flash card, and would do an export and then an import, the import would change it to 读 ⁵d?u. If you look at the export ed file, it would still be saved as 读 d?
- if you export/import a word that has a "special character" in it, after importing it the meaning will no longer show when being tested. To examples that I have seen are: 希/稀奇 xīq? and 花(儿) huā(r)

- if you have the % Correct shown in the flash card session, it will alway give a result of 0% or 100% (am like it is being treated as an int)

- the testing history does not seem be recording accurately (not recording everytime whether it is correct or not). I had a list of two cards that I tested a number of times and it would not record the incorrect entries, but would record the correct entries.


Well, those are the bugs for now. I really love the additions that you have added to the flashcard portion of the dictionary. I find them a night and day improvement over the older version.

Lonnie

 

[pete17]

Hey Lonnie,

Nice bug list, well thought out ... As you know, I'm still struggling to get the T5 to recognize the ABC dict on its internal mem ... I saw a post somewhere on this Forum, someone actually had the courage to autoload PlecoDict on the T5, but I may just have to wait until the next release ...

- Pete

 

[Anonymous]

A few observations

Just installed PlecoDict on my Zodiac 2, and it's great!!!

*** I have nothing to do with Tapwave, but I have to plug the Zodiac ***
First: Anyone thinking about what Palm hardware to get should consider the Tapwave Zodiac 2! It's marketed as a gaming machine, so you may not expect it to be such a great Palm device. Metal case, 128MB RAM, two SD slots, beautiful screen, ... I've had mine for a month and I love it.
******************

Here are a few observations:

1. My first try installing resulted in this HotSync error:
ERROR: The following file(s) could not be installed to the Internal File System because there is no application on your handheld to open these files.
If you have recently installed such an application, please run that application and then perform a HotSync operation.
NWP.pdb
Oxford C-E.pdb
Oxford E-C.pdb

So I manually installed these files to "handheld" instead of "internal file system" and all is well.

2. I'm seeing the problem with Roman numerals that Henry pointed out. The bolding is inconsistent, and the numerals are too close to the preceding word, so there appear to be lots of pinyin words ending in lower-case "ell".

3. Using the big font size, the Chinese characters are too close together, and adjacent characters often touch. It's subtle, but it has a big impact on readability - the separate components within a character are separated more than they are from parts of the characters before and after.

4. The pulldown arrow for recent entries is less convenient now that it's on the right instead of the left, since the left side is where the entry appears. Since I'm using PlecoDict in landscape mode (480x320), this is especially noticeable.

 

[mikelove]

beirne - you're right, it only displays the alternate with the single-character entry. Odd. We really don't have the scholarly resources here to resolve this question, but we do have the rights to updated versions of the ABC dictionary, so at whatever point they resolve this we should hopefully be able to work those changes into PlecoDict.

Ole - strange, must be a bug in the text converter. (perhaps it's mistakenly treating the umlaut as part of a Pinyin syllable)

Lonnie - the missing cut/copy/paste icons are a simple omission on our part, since we don't use the built-in Palm OS system for text fields we have to add them manually and we just haven't gotten around to it yet.

We've gotten a few reports of those flashcard bugs you mentioned, the manage lists one is again just a bit of laziness on our part and the frequency adjustment one is related to a number of stability problems with new flashcard sessions. (on some Palms they refuse to work altogether)

The import problem is just a weakness in our text importer: at the moment, it only searches for matching characters and doesn't attempt to narrow them down to only those that match on Pinyin as well. Of course, even if it did it still wouldn't work with characters that had multiple definitions with the same pronunciation... the Pinyin problem should be fixed in 1.0, we're planning a more robust import/export system for 1.1 or some future release.

That's strange about the incorrect entries not being recorded - was there any other pattern there? For example, did you always mark the first card correct and the second card incorrect, or something like that?

pete17 - see my recent e-mail on this, assuming you put ABC.pdb in /Palm/Launcher/ or /Palm/Programs/PlecoDict/ it *should* be able to recognize it.

Steve - agree with you on the Zodiac (see http://www.pleco.com/whichpalm.html where we mention it quite positively). As for your observations:

1. We haven't gotten any reports of this sort of bug before - it might be something specific to the Zodiac's version of HotSync Manager. Normally this error comes up when you try to HotSync a file to VFS and Palm Desktop doesn't recognize the format (e.g. a Word document or an unusual image file) - every file installed by a HotSync has to have some application that knows how to read it. But it shouldn't be coming up for PDB files.

2. Yeah, one of the many things we're working to fix in the not-quite-perfectly-converted Oxford data files.

3. Another data file problem, this time with the large font file (some other people have already pointed out flaws in a few of the characters) - thanks for reminding me of this, though, it definitely does have a negative impact on readability.

4. We actually thought that it would be better in the new position, because it makes the Input Field easier to reach (all you have to do is tap the edge of the screen) - does anyone else feel it was better off in the old location?

 

[pete17]

Mike, I moved everything into a Palm/Launcher folder (which for some reason had not been created automatically) and I've got ABC dict working!
Thanks for tracking along with me. It's a terrific product! I particularly like the resolution on the T5 screen.

- Pete

 

[Anonymous]

Mike,

You said: "4. We actually thought that it would be better in the new position, because it makes the Input Field easier to reach (all you have to do is tap the edge of the screen) - does anyone else feel it was better off in the old location?"

The pulldown arrow I'm talking about is what you call the "Recent Query List". It was at the edge of the screen before (the left edge), and now it's not. See http://pleco.com/images/screenshot.gif. You describe it in the manual this way "Below the toolbar is the Input Field - this is where you enter text to look up in the dictionary. Next to that is the Recent Query List - tap on that downward-pointing triangle to bring up a list of words you've looked up recently. And to the right of that is the Clear Button, which clears the contents of the Input Field." Since there's something to the left and something to the right, I'm not sure which edge of the screen you're talking about.

A couple other things about the Recent Query List:

The back and forward buttons make my gripe a lot less important - I suspect I'll end up using "back" more if the pulldown stays where it is.

The "Switch List Mode" option seems to work differently than it used to. It either does nothing (if I'm not looking at search results), or it toggles between a "list" of results, usually containing just one word, and a list of many. In Oxford, it did this just once, then went to a default browsing mode where it toggled between an alphabetical list of single characters around the current entry and a list of words around the current entry. If I'm reading the dictionary entry for da4, Oxford toggles the list between (da4, da4bai2, da4bai2cai4, da4ban4, ...), and (da4, dai1, dai1, dai3tu2, ...) which I find useful (these lists scroll up to see things before da4). Does this make sense?

I *love* the fact that I can use small font for the list and large font for the main definition area!!

Being able to keep the handwriting recognition feature open and docked is also a real treat.

 

[mikelove]

I was actually referring to the Input Field being at the edge of the screen; we wanted to make it as easy as possible for people to tap on that. We might consider making this a Preferences option in version 1.1, though.

The changes to Switch List Mode were designed to make the interface more straightforward while still providing most of the same benefits of the Oxford E&C list modes. For example, you still get the single-character list when you type in a single-character (or syllable) query. Could you be a little more specific about when you find the ability to scroll up past the first result (or to quickly switch to single-character mode for a multiple-character search) useful? There's not really any technical reason we can't do this, we just don't quite see the point.

I'm glad you find the differing font sizes option useful - widescreen devices like the Zodiac were one of the main reasons we added it, since you've got plenty of screen area but not very much vertical size for list entries. The docked handwriting recognition is also a large-screen thing.

 

[Anonymous]

I was actually referring to the Input Field being at the edge of the screen; we wanted to make it as easy as possible for people to tap on that. We might consider making this a Preferences option in version 1.1, though.

I see. I've never though of a tap location near an edge as being easier. For me, having related things near each other is what's important - The pulldown and the cursor are two ways to fill the input field, and they have been moved apart in Pleco.

By the way, in PlecoDict, when the input field is empty, I do not see a cursor, so I end up tapping there when I don't need to. Once there is at least one letter or character there, the cursor is back. Is this a Zodiac thing?

The changes to Switch List Mode were designed to make the interface more straightforward while still providing most of the same benefits of the Oxford E&C list modes. For example, you still get the single-character list when you type in a single-character (or syllable) query. Could you be a little more specific about when you find the ability to scroll up past the first result (or to quickly switch to single-character mode for a multiple-character search) useful? There's not really any technical reason we can't do this, we just don't quite see the point.

Ok - I've discovered what's going on. In Oxford, there are three views (I never noticed that). In Pleco there are only two. I miss the Character mode view - that's the main issue, since I used it and the pinyin view both for browsing the dictionary (character view let's me browse faster, then I'd toggle to pinyin...). How the lists center themselves is what I was getting at before, but it's a can of worms, and if there is only one full dictionary view, it's a moot point.

I think what I want is three list views: character, word, and search-results, the first two of which would always be real windows into the complete dictionary. In Pleco, we only have search results view and word view and so there's no character view for scrolling through the dictionary, and in Pleco, switch view sometimes doesn't do anything - if there's nothing searched for.

Steve

 

[goulniky]

One small bug I just came across is :
I select a character in ABC mode, handwritten recognition, say chen4 衬. I get a definition, it ends with chen4yi (i.e. see chen4yi). I select and "copy to input field" the last syllable (yi), expecting I would get a mixed input field, character (衬) followed by pinyin (yi).
What happens instead is Yi gets copied to the input field (erasing the character) and I get the first Yi in the list, but the dictionary icon switches from ABC to NWP for some reason.

 

[mikelove]

SK - interesting point about the edge location, and I suppose it is more normal to have the arrow at the left of the Input Field instead of the right. The cursor problem is universal, we actually use a system of our own design for text fields (the built-in Palm OS one doesn't offer the features we need) and we've had a hard time wrangling the cursor into behaving the way it's supposed to. And I see your point about single-character mode; it's quite impractical to browse through a 200,000-entry dictionary word-by-word. But for that to really be useful we'd also have to get scroll bars working for entry lists, which would mean writing our very own version of the Palm OS scrollbar (the built-in one won't accommodate the number of entries we need - with the Oxford it's OK but the ABC is just too big) - it would certainly be a nice feature but it's not going to happen in 1.0.

goulniky - this isn't really a bug so much as a poorly-implemented feature; Copy to Input Field always assumes you're trying to look up a brand new word (rather than just part of one), and it thinks Pinyin is English so it tries to switch to the English-to-Chinese dictionary whenever you run it on Pinyin. There should probably be a preferences setting for whether or not to clear the Input Field, and the language detection really needs to be more intelligent about Pinyin. Both of these things should be corrected in a future release, but for now I'd recommend disabling the "Auto-detect Copy to Input lang." setting in Preferences and using copy-and-paste when you want to insert the second syllable into a word.

 

[shindong]

pleco dict with pocket tunes?

hi mike.

I using T5 and a few weeks ago had asked about pleco.DB file replacement error message.

as you recommended I moved all font files and dictionary data files from 'internal' to 'INTERNAL'.

firstly the crash in the middle of search problem seems fine now.
it does not crash much. (but still sometimes shows error message about free memory)

but not like you mentioned before, searching speed is much slower than before.(when all the file was in 'internal')

many times even handwriting graffiti of Palm doesn't follows my stroke. (and.... sometimes Pleco Dict's chinese full screen handwriting recognition as well)

everyday.. it gets slower and slower..
so I have to do soft reset anyways.

I think it is related with T5's new memory. (especially virtual hip(?) memory things for run program on RAM)
concerning about returning hip(?) memory problem posibility. (or leaking memory..)

I using "pocket tunes 3.0.3" to set "Parmary Clock"'s MP3 alarm sound option enable.
and I think these two pretty heavy application consuming it's memory out. (just my guess)

hope these slower problem might be solved for Beta2 version already.

 

[mikelove]

That's odd - there really shouldn't be much of a difference. The handwriting slowdown doesn't make any sense; nothing about where the data files are stored has any impact on the handwriting system's stroke capture. We haven't encountered any of these slowdowns in our testing (and we don't know of any memory leaks at present), but we'll keep looking and maybe try some experiments with PocketTunes to see if we come up with anything.

Have you installed the latest software updates for your T5? The last one (dated October 24th) is available at http://www.palmone.com/us/support/downloads/tungstent5/tungstent5update.html.

 

[shindong]

more detail..

thanx for reply..
I've checked T5's latest update in the link that you attached.
firstly I'm using chinese version. (I bought T5 in china)
and.. version seems latest.

I also think pleco dict doesn't affect on Palm's hand writing recognition system, but pynin search and display was slow anyhow.

I'll give you some more detail.
hand writing stroke problem actually seems..
pynin search speed don't follows hand writing recognition.
when I finished first character's stroke, there are character that I wrote in input field.
But, if I write next character with same speed that I usually write on memopad.. on screen there are no stroke trail line and no response.
right after the fisrt character's stkoke recognition finished, I can see displaying delay.
it refreshs from top to bottom.. and it takes about 0.5s at least.
if you say 0.5s is short time and I can wait.. nothing to say.. but when data file was in 'internal'.. it wasn't like this slow.

anyway..
now I deleted some not desperately needed mp3 files and made more free space on 'INTERNAL' and did soft reset.
(before time there was 9.9MB free memory on 'INTERNAL' and now have 23MB)
I seems better.
I'm still not sure about the gettin' slower problem completely solved..
if there are any significant, I will report! ^^

加油！

 

[mikelove]

Hmm... well the best way to deal with the software being slow is to enable the "delay before searching" option in Preferences; that way, the software won't begin searching until a certain amount of time has passed since you last entered a character, so you won't encounter any holdups while you're entering characters.

Also, as a compromise on speed you could try moving the font files to Internal while keeping the dictionary data files in INTERNAL - the slow screen refresh is likely a font-related problem.

Anyway, thanks for the feedback on this, and I'm sorry you're having so many problems...

 

[Jianchuanbrit]

font problem

I know that fonts are still being worked on but I just noticed something in the Chinese font that bothers me so thought I would bring it up here.
The character 'fang' as in house has a dian at the top, but it is displayed as having a heng across the top. It is shown like this in the input line, head entry and scroll list on right hand side. If you tap on it the enlarged font displays correctly. I think it might confuse some people, and I would have thought it should be possible to display characters like this correctly. Hope it can be fixed.

 

[mikelove]

This has already been fixed in Beta 2 (which should be available on the 15th, though that's just a guess and not an official statement). (technically that's still "early April" if you define "early" as "first half of" )

 

[Henry]

add to flashcards

If I am not looking closely it is sometimes hard to tell if plecodict has registered my quick tap on the 'add to flashcards' button, would it be possible to make it a bit more obvious when it is successful? (I have no good ideas . . .)

I'm happy to see that when I have tapped the 'add to flashcards' button several times only one copy of the entry is added!

Henry

 

[mikelove]

A couple of people have mentioned this, but we haven't been able to come up with any good ideas either, which is why we added that duplicate-eliminating feature. Animations are a problem since the Palm OS is not multithreaded - any animation would likely block most other user input until it finished, and even a half-second delay could get very annoying when entering lots of words.

 

[Smoodo]

I have an idea to solve the dup addition problem

You know that window that pops up and zooms in on a character? How about you have that pop up with the word "added" under it and zoom in on the character(s) that actually made it into your flashcard file when you click the add to flashcard button.

This should be easy to add, within the confines of your input scheme, and doesn't require doing anything tricky with icons.

How about it or some variation?

 

[mikelove]

That should work, but we're really really trying not to add any more features (even small and useful ones) at this point - it's a slippery slope and we just can't afford to keep delaying the finished version. If we can confidently do this without creating any bugs, though, we might give it a try...