2.2.11 Bug Report / Feedback Thread

[radioman]

I'd tend to agree, though we are adding the option for a dedicated dictionary switch button to appear once a flashcard is revealed.

I'm understand there is logic to the way that it is currently set up. My position is that I do not want to have to go and seek out a tiny little button all the time when I'm running through flashcards, or Im on a dictionary lookup. I swap dictionaries quite often and if there is 50-80% of free space on the screen, I would like to have the ability to press some of the empty space to swap. Then if the dictionaries are prioritized, the next lookup or flashcard resumes at the top of the prioritization list.

I do not think this type of "macro-gesture" option is any more unusual than tapping the reader sides to go to the previous or next page. Perhaps there should be a more broad option selection for the free space area, so that someone can select swapping dictionaries, playing audio, etc.

Resize so that the width is set to the width of all recognized characters?

Yes, that is what I am thinking. However, if paragraphs are offset, maybe there could be a way to resize to the largest recognized character area for the place where the screen is tapped.

 

[mikelove]

I'm understand there is logic to the way that it is currently set up. My position is that I do not want to have to go and seek out a tiny little button all the time when I'm running through flashcards, or Im on a dictionary lookup. I swap dictionaries quite often and if there is 50-80% of free space on the screen, I would like to have the ability to press some of the empty space to swap. Then if the dictionaries are prioritized, the next lookup or flashcard resumes at the top of the prioritization list.

It's just not a very intuitive gesture - tap in a part of this screen that happens to contain text to cross-reference the text, tap in a blank space to do something else; a user has to consciously think about the fact that they need to look for blank space / find it / tap in it to invoke the gesture, or get in the habit of always tapping in a certain space only to have it not work on the occasions when that space contains a character. If a gesture requires as much of a mental burden to invoke as a dedicated button does, if it's not something that you can just do instantly without thinking about it, then it really becomes hard to justify supporting it.

 

[character]

My position is that I do not want to have to go and seek out a tiny little button all the time when I'm running through flashcards, or Im on a dictionary lookup.

It would definitely help muscle memory if there was a dictionary switch button in the same place on both the dictionary and flashcard views. I don't know if that was a consideration as part of the upcoming redesign.

Thanks Mike for the information about how definitions for flashcards are chosen. I guess the behavior I've seen has been because of early on importing a pre-built HSK flashcard set which must have had its own definitions, so those are overriding the dictionary definitions.

 

[radioman]

This one fits with your earlier OCR suggestions I think - some sort of improved resize behavior might happen, but the main thing we're trying to get added to OCR in the next update is PDF support.

Yes, that is correct. But to be clear, my main interest is far and away getting the PDF support. The PNG references are by and large just a work-around for the grander PDF support. That is, any references to OCR features I may have made (e.g., double-tap a PNG file) are features I think would be first and foremost useful for for the PDF function.

I know the PDF development is well down the road. But should it be of use, I wanted to provide my recent observations with regard to the PNG OCR document workaround. I figure they can be attributed to my unorthodox use of the OCR.

1) If I have nine (or only two) PNG files concatenated into one file, I cannot mark across the multiple pages (sub-pages) that make up the PNG file and successfully OCR. I have to set the OCR box around the one sub-page.

2) Depending upon the makeup of a given set of sub-pages, the OCR result might be clean on a per sub-page basis. After OCR, I typically enlarge the text of a sub-page so that it fits across the entire screen width in landscape, which effectively means that whatever is on the screen (and beyond) is OCRed. However, because of the current limitation with OCRing more than one sub-page at a time, I see value in a mode where anything in the screen is auto OCRed (and nothing else). So in my current configuration, it would then be just a simple matter of sliding to the next page.

I believe this field-of-view mode was previously referenced on the forum (maybe by @character). It would obviously not eliminate the need to be able to flexibly configure the OCR border box, it would just be a convenient reset-the-box mode. But at least with regard to speed of moving between sub-pages, and re-OCRing adjacent sub-pages, Pleco seems plenty fast on the iPad 2.

*** Pleco PDF OCR - Version 2? ***
I see these applications in iTunes and for Android where you have the video of a street sign in some language and the app presents you with the translated text in the same font, so it looks like the original sign, but in a new language. Perhaps Pleco could do just that but within documents, where the PDF paragraphs or pages would magically be translated and rendered on the fly. I know that the translations are an imperfect science but they already useful, and getting better. I think that would be really cool to be able to take a Chinese document, flip it via some English Translation engine to get the gist, and then go back to the Hanzi view to start the serious study effort. Or similarly, transliteration to Pinyin, Chinese > English, etc.

This gets back to the idea of some business guy unskilled in Chinese language sitting in some meeting having to sit through a bunch of Hanzi-based Powerpoint presentations. However, even for the Chinese student, being able to quickly get the gist of a 5 page document before plowing into it (or quickly reviewing a difficult 4-line paragraph's meaning) prior to personally deciphering the Hanzi would be useful (I know it would be helpful to me ).

There are document translation services out there (e.g., http://goo.gl/qOSa ) but thats not "on the fly" and in a nice, integrated environment, where a paragraph can be magically flipped in place.

Maybe I am wrong, but I am thinking that since Pleco OCR works directly with the graphic files (extracted from a PDF or otherwise), it puts Pleco in the unique position of being able to deal directly with the graphics, graphically overlaying text, etc, exactly where it would need to placed and at the proper font (rather than having to fit it into some sort of text-based form) to make it look right.

够了

 

[mikelove]

I know the PDF development is well down the road. But should it be of use, I wanted to provide my recent observations with regard to the PNG OCR document workaround. I figure they can be attributed to my unorthodox use of the OCR.

Oh it's not well down the road at all, as I said it's already working in our test builds here. If it weren't for the pending release of the new iPad and the iPhone 5 we might already be doing a limited iPhone beta-test (or at least gearing up for one), but thanks to those we can't really afford to hand out a bunch of our precious 100 test device slots now when people are about to turn right around and replace them (and when we need to recruit people with those replacement devices).

I see these applications in iTunes and for Android where you have the video of a street sign in some language and the app presents you with the translated text in the same font, so it looks like the original sign, but in a new language. Perhaps Pleco could do just that but within documents, where the PDF paragraphs or pages would magically be translated and rendered on the fly. I know that the translations are an imperfect science but they already useful, and getting better. I think that would be really cool to be able to take a Chinese document, flip it via some English Translation engine to get the gist, and then go back to the Hanzi view to start the serious study effort. Or similarly, transliteration to Pinyin, Chinese > English, etc.

That really doesn't work well - have you actually tried out those apps? Most of them can really only translate word-by-word - the picture of the text is too unreliable and online translation software too slow to do more intelligent sentence translation - and with Chinese it's especially tricky because you need the entire sentence in order to even figure out where the boundaries between words are. (there's also the practical issue that English translations of Chinese text tend to take up a lot more space - you'd have a very hard time squeezing it in without using insanely skinny fonts)

More generally, though, we've made the conscious decision not to focus on online translation and we stand by that. We're never going to be able to compete with companies like Google and Microsoft (both of which have been moving towards offering this exact same technology in their own apps - it's only a matter of time before Google Goggles lets you snap a picture of a page of Chinese and spit out a Google Translate English version, and does all of that for free) and it's not what we're about anyway; we make software to help people with their Chinese, not software to make Chinese accessible to people who don't know it at all.

Maybe I am wrong, but I am thinking that since Pleco OCR works directly with the graphic files (extracted from a PDF or otherwise), it puts Pleco in the unique position of being able to deal directly with the graphics, graphically overlaying text, etc, exactly where it would need to placed and at the proper font (rather than having to fit it into some sort of text-based form) to make it look right.

We certainly could - it's not a difficult effect to mimic, really, we already know the character positions and the colors are quite easy to figure out - but again, it's not an area where we think we can make any money because there are too many companies out there already giving this away for free or about to start giving it away for free. So it's not just worth it - boutique software firms like Pleco have to focus on a) things that bigger companies aren't interested in or b) things that bigger companies won't automatically be better at, and this doesn't meet either of those requirements; it's a huge potential market, and Google and Microsoft can easily spend their way to a better OCR system than ours.

 

[character]

I see these applications in iTunes and for Android where you have the video of a street sign in some language and the app presents you with the translated text in the same font, so it looks like the original sign, but in a new language.

I have zero interest in this, and strong interest in OCR quality improving. Right now I don't trust OCR enough to use it to extract the characters from an image and dump them to the reader. Pleco's OCR is magic compared to not having OCR, but it's still not where it could be (take a reasonable picture of a page in normal lighting from a standard text-only book and be confident that the text was correctly recognized without recognition errors, extraneous characters, etc.).

Also really don't want to see Pleco dependent on a network connection.

 

[mikelove]

I have zero interest in this, and strong interest in OCR quality improving. Right now I don't trust OCR enough to use it to extract the characters from an image and dump them to the reader. Pleco's OCR is magic compared to not having OCR, but it's still not where it could be (take a reasonable picture of a page in normal lighting from a standard text-only book and be confident that the text was correctly recognized without recognition errors, extraneous characters, etc.).

For that we're probably going to need to license a different recognition library / algorithm, one that's much slower and hence not suitable for live video but better at still images (really more of the sort of algorithm you'd use on a regular PC with a scanner); that's something we're currently investigating. In the meantime we are working on improvements to live video accuracy / ergonomics, and on adding PDF support to still images.

Also really don't want to see Pleco dependent on a network connection.

No plans for that, don't worry.

 

[character]

For that we're probably going to need to license a different recognition library / algorithm, one that's much slower and hence not suitable for live video but better at still images (really more of the sort of algorithm you'd use on a regular PC with a scanner); that's something we're currently investigating.

Great news; would happily pay again for 'Still OCR' or whatever, esp. if it also was great with Traditional characters.

 

[radioman]

OCR ****
Is there anyway to take an OCRed page with green still on it, re-OCR a character via press-and-hold, and just delete that character?

The application for this is so that if you have a sentence with some reference designator (e.g., [j] ), where
- the reference is typically OCRed wrong because it uses unusual characters or offset from the rest of the text, and
- it does not help you decipher the context of the sentence.

After cleaning up a number of these designators on a page, clipping the page will provide a clean result.

URL Call ****
This might be more of a feature request than anything. But, referencing the URL call function to Pleco, such as the one Skritter uses, Is it possible to do a URL call to initiate Pleco:
1) into the web browser function, and/or
2) have a specific URL included in the URL call string to bring up a specific webpage?

 

[mikelove]

Is there anyway to take an OCRed page with green still on it, re-OCR a character via press-and-hold, and just delete that character?

Not at the moment, but it's a good idea.

This might be more of a feature request than anything. But, referencing the URL call function to Pleco, such as the one Skritter uses, Is it possible to do a URL call to initiate Pleco:
1) into the web browser function, and/or
2) have a specific URL included in the URL call string to bring up a specific webpage?

Possible, but it's tough to justify adding these things unless we have another popular / likely-to-be-popular app (like Skritter) wanting to use them, since they actually take quite a bit of effort to get working well.

 

[radioman]

I have been getting a number of failures lately with regard to registration, where the registration data is lost and I have to re-enter the code. Not sure if this is something that has been reported in the past, but if there are any ideas on correcting this, please pass them along.

EDIT: Not sure if this might have to do with some sort of checks via VPN versus registration via China networks.

 

[mikelove]

I have been getting a number of failures lately with regard to registration, where the registration data is lost and I have to re-enter the code. Not sure if this is something that has been reported in the past, but if there are any ideas on correcting this, please pass them along.

EDIT: Not sure if this might have to do with some sort of checks via VPN versus registration via China networks.

Have you done anything that might wipe private files / data on your system?

 

[scykei]

Hey, can you check if you can properly unzip compressed zip files with the file browser on iOS? I have no idea why but I cannot seem to get it to work. I have tried importing a zip archive of several plain text documents in several layers of folders using USB and Wifi but neither seems to work at all. I have even tried to email a copy to myself and have the document open in Pleco from there. Nothing seems to work. :\

I recently had a problem with my iPad 3 and after sending it back to Apple and waiting for a week, I got a new one in return. I had a lot of documents saved up in there. Luckily I had a backup on my computer but I just can't transfer it to my iPad in bulk. I have not tried transferring a zip file to Pleco before so I'm not sure if it was ever possible to do in the first place.

What I notice is that the file doesn't completely get transferred into Pleco. Pleco will show the size of the zip file after transfer, but when I decompress it, it just makes an empty folder. I've tried zipping it with zero compression with both WinRAR and 7-Zip.

What am I doing wrong? :\

 

[radioman]

Have you done anything that might wipe private files / data on your system?

Well, I was moving docs in and out of the Documents directory in iOS directly. And I think there are hidden files and directories there that store various settings.

I know I am not deleting those files, but maybe I screwed up the permissions, leaving some of the files susceptible to being overwritten. I have no such problem on my iPhone, where I am not moving copious amounts of files in and out.

Sorry. I know this stems beyond what you support. Worse comes to worse, I'll just do a full reinstall.

 

[mikelove]

Hey, can you check if you can properly unzip compressed zip files with the file browser on iOS? I have no idea why but I cannot seem to get it to work. I have tried importing a zip archive of several plain text documents in several layers of folders using USB and Wifi but neither seems to work at all. I have even tried to email a copy to myself and have the document open in Pleco from there. Nothing seems to work. :\

They work fine for me - which specific zip settings are you using in WinRar/7-Zip? Perhaps there's some aspect of that spec that Pleco doesn't support.

Well, I was moving docs in and out of the Documents directory in iOS directly. And I think there are hidden files and directories there that store various settings.

I know I am not deleting those files, but maybe I screwed up the permissions, leaving some of the files susceptible to being overwritten. I have no such problem on my iPhone, where I am not moving copious amounts of files in and out.

That could mix up the registration, yes. It could also be that you're overwriting a folder that contains hidden files with another folder that lacks those files.

 

[scykei]

Hey, can you check if you can properly unzip compressed zip files with the file browser on iOS? I have no idea why but I cannot seem to get it to work. I have tried importing a zip archive of several plain text documents in several layers of folders using USB and Wifi but neither seems to work at all. I have even tried to email a copy to myself and have the document open in Pleco from there. Nothing seems to work. :\

They work fine for me - which specific zip settings are you using in WinRar/7-Zip? Perhaps there's some aspect of that spec that Pleco doesn't support.

I've done some investigation. Basically, I tried creating new sample txt files, zip them and upload them to pleco to test. This works.

But certain documents when downloaded from the internet will not. I'm not sure what the problem is but they just don't work. So, I've taken a few files, removed most of the content leaving minimal data in it. I'll upload them here for you to check them out.

When in a folder, it just extracts it out without any content. Without, after extraction it creates a weird hidden file. I'm not sure why, so I made two zip files. I used the exact same settings with the ones I could make work.

Thanks.

 

[radioman]

Possible, but it's tough to justify adding these things unless we have another popular / likely-to-be-popular app (like Skritter) wanting to use them, since they actually take quite a bit of effort to get working well.

I understand the need for prioritization. But I would like to expand on this just a bit more.

First, Pleco Reader is truly unique. Maybe I'm wrong, but to me, this just seems like a "build it and they will come" scenario. The Reader is just too useful. I believe there are any number of publications, applications, etc. that would be willing to support this. I currently utilize a number of programs daily to drive content via the pasteboard into the Pleco Reader, from Chinese News Applications, to the program "Drafts", and others. The pasteboard function is certainly useful as is. But getting the round-trip button onto the reader in this way would make the overall presentation less laborious and more controlled.

The second idea revolves around being able to start Pleco up in the various modules (e.g., Dictionary vs. Reader vs. Flashcards, etc. ) and have it user selectable. Assuming Pleco to remain as an integrated set of tools, I was thinking that if the URL call to initiate Pleco could also specify the module and content when starting or being activated, then there might be a way to save this URL call via Safari and write it as a bookmark springboard icon. In this way, the icons woud look like separate programs and provide easy access to frequently used Pleco functions. (e.g., Dictionary, Reader to a specific document, Pleco Browser to a specific Chinese news webpage).

 

[mikelove]

When in a folder, it just extracts it out without any content. Without, after extraction it creates a weird hidden file. I'm not sure why, so I made two zip files. I used the exact same settings with the ones I could make work.

These seem to work fine for me - were the names of the files different on your device by any chance? What version of iOS (and of Pleco) are you using?

First, Pleco Reader is truly unique. Maybe I'm wrong, but to me, this just seems like a "build it and they will come" scenario. The Reader is just too useful. I believe there are any number of publications, applications, etc. that would be willing to support this. I currently utilize a number of programs daily to drive content via the pasteboard into the Pleco Reader, from Chinese News Applications, to the program "Drafts", and others. The pasteboard function is certainly useful as is. But getting the round-trip button onto the reader in this way would make the overall presentation less laborious and more controlled.

I see your point, but we've got a long list of equally-justifiable new features with more widespread user interest - Skritter was an exception because it had to happen right away and we knew it was an app that a lot of our customers like / would like, but I don't know of any other app right now with the same potential appeal.

We do, however, want to improve our "View In..." support to spare users from having to go into Inbox to view documents, so that's one area we may be able to help on this - would be a more standardized way to do it anyway.

The second idea revolves around being able to start Pleco up in the various modules (e.g., Dictionary vs. Reader vs. Flashcards, etc. ) and have it user selectable. Assuming Pleco to remain as an integrated set of tools, I was thinking that if the URL call to initiate Pleco could also specify the module and content when starting or being activated, then there might be a way to save this URL call via Safari and write it as a bookmark springboard icon. In this way, the icons woud look like separate programs and provide easy access to frequently used Pleco functions. (e.g., Dictionary, Reader to a specific document, Pleco Browser to a specific Chinese news webpage).

We actually did that on Android - created separate Flashcard / Reader / OCR launcher apps - but then a competitor accused us of using those launcher apps to spam Android Market and threatened to report us to Google (and we had reason to think this was a serious threat), so we got rid of them. Which wouldn't bear on this except that in the month or so since we got rid of them we've received exactly 1 complaint about it, leading me to think that this maybe isn't something with much widespread interest either. And on iOS there'd be even less interest because it would be harder to use - there's no chance of Apple approving an app that consists solely of a shortcut to another app, and there's no way to easily generate home screen shortcuts in-app.

 

[scykei]

When in a folder, it just extracts it out without any content. Without, after extraction it creates a weird hidden file. I'm not sure why, so I made two zip files. I used the exact same settings with the ones I could make work.

These seem to work fine for me - were the names of the files different on your device by any chance? What version of iOS (and of Pleco) are you using?

iPad 3 iOS 6 (Although it never worked on iOS 5 either)
Latest Pleco running 2.2.11.

So I'm the only one getting this? I've decided to record a video of this. The best thing I had for recording video was an iPod Touch 4 so the quality wouldn't be too great.
http://www.youtube.com/watch?v=2bNtDAPexf0

 

[radioman]

there's no chance of Apple approving an app that consists solely of a shortcut to another app, and there's no way to easily generate home screen shortcuts in-app.

The approach I was envisioning was more or less a hack, just an external option listed in the user manual or on a website. If someone wanted to set up the shortcuts on their own using Safari, they could do so. (perhaps some way to just download the links on some external site). They would need to put in the links and then save them to the springboard.

Selling springboard shortcuts through the app store I'm sure would be another matter, and not one I was contemplating.