iPhone Feature Requests

[gato]

I was actually thinking of legal ones - old series that are available for free, or archives of subtitle files that can be used with cheaply-available DVDs. My sense at least is that there are a great many places online where one can perfectly legally watch Chinese TV for free; is that not the case?

Many P2P video download sites have been shut down in crackdowns in the last two years. But you can watch many movies and TV shows for free online at sites like qiyi.com or tudou.com. They have all gone legit and pay to license their content.

http://www.variety.com/article/VR1118030821?refCatId=19
VeryCD website suspends services
Move may signal crackdown on copyright breaches in China
By Clifford Coonan
BEIJING -- Free music and movie download services at one of China's most popular download sites, VeryCD, were blocked in recent days in a move that could signal a crackdown on copyright breaches in China.

Access to music channels and movie clips at VeryCD, which registers up to three million hits a day, have either been blocked or automatically diverted to other web pages since Saturday.

Downloading functions on other smaller peer-to-peer file-sharing sites have also been suspended.

Local media in China has reported that more than 400 peer-to-peer file-sharing sites, including Btchina, China's top media file-sharing website, have been shut down during the last two years.

 

[mikelove]

I have no problem with having pinyin as the source for the pronunciation so perhaps adding another field that can be populated by the dictionary. Supporting all the other romanization systems seems like a headache! GR Itself has so many weird exceptions and spelling rules, especially between putonghua and peking dialect. I know I'm in the minority using GR so wouldn't expect it to be explicitly supported. It'd just be nice if I could use my flash cards for listening practice. Also the goal is at some point to just completely abandon romanization systems and move to characters.

Well a "custom fields" system wouldn't be too difficult to add and would let you have both simultaneously while also adding other helpful things like usage notes - that's probably what we'll end up doing instead of just mixing pronunciations. (though we'll have to provide some sort of migration path for @ users)

I did a test using Pleco OCR on a screencaptured image of the video playing with subtitles. As you can see, though the subtitles are pretty clear, the chars are far too small for the OCR to guess them...if there's a way to blow up the image and give the OCR bigger characters to look at, I didn't find it.

It looks like the problem here is that the image is rotated and zoomed-out; fix the first problem by tapping on the middle button in the top toolbar, then on the rotate arrow, and fix the other problem with a pinch-zoom (along with resizing the box to cover just the characters) and that should give you a better idea. The resolution of most subtitles should be fine for Pleco, the main worry is whether or not there's a distracting background / color borders that might confuse the recognizer.

Here is what I would love to see... more sentence examples, specially in flashcards.

Let me elaborate.... ABC is in that respect already quite good. But some structure words really can be very tricky. A simple noun like 狗 will be no problem, but 当 or even the simple 对 can have different meanings.

So what I think is that a sentence DB would be a good thing, were you have i.e. different examples of 当 in a sentence. This DB could be even user generated.

This help Pleco as a learning tool. I think there is no doubt that one needs sentences to get the characters into the passive memory. This could be a good step.

We've got a lot more in PLC than in ABC - are those still not sufficient? Tuttle is also full of them, and a few of the new dictionaries we're launching this summer have a bunch as well.

The sentence database idea is something we played around with for a while, but there's a big problem we couldn't come up with a good answer for: unless you actually tag sentences with the "sense" of the word as well as the word itself, it can be very difficult to figure out which of the many different meanings of 当 the sentence relates to. This is why example sentences in dictionary definitions are so valuable; the example illustrates the sense of the word that the dictionary just described. So you're getting a lot more out of seeing it right there in that context than you would seeing it as part of a long list of undifferentiated sentences.

Another issue, but it would be fantastic if during a flashcard session, the example sentence could be read out by some kind of text-to-speech function inside Pleco or Google Translate. Pleco's current audio pronunciation function is wonderful when doing single words but not so great when doing sentences. I use another flashcard program called Anki for my Japanese and the version for Android called Ankidroid has just recently added this TTS functionality, and while they don't have it perfect yet, I think it's a good example of this.

Look for an announcement on this front very soon

 

[mikeo]

I did a test using Pleco OCR on a screencaptured image of the video playing with subtitles. As you can see, though the subtitles are pretty clear, the chars are far too small for the OCR to guess them...if there's a way to blow up the image and give the OCR bigger characters to look at, I didn't find it.

It looks like the problem here is that the image is rotated and zoomed-out; fix the first problem by tapping on the middle button in the top toolbar, then on the rotate arrow, and fix the other problem with a pinch-zoom (along with resizing the box to cover just the characters) and that should give you a better idea. The resolution of most subtitles should be fine for Pleco, the main worry is whether or not there's a distracting background / color borders that might confuse the recognizer.

Yes, I agree that with enough rotation changes, resizes, and refocuses you could probably get OCR to work on these subtitles.

But unless a lot of these steps could be automated, to resize the text large enough for OCR, and allow the user to easily select it and get an immediate lookup , the whole process would be a whole lot slower and more tortuous than simply doing a pinyin lookup.

Usually, it's a single phrase or character in the subtitle you want to look up ( if you need to look up every character or phrase in almost every subtitle, you'll give up soon enough) . Also, I don't think Pleco is designed to do translation of whole sentences or phrases, such as appear in subtitles, so you'd probably not have it doing automated subtitle translation.

 

[mikelove]

But unless a lot of these steps could be automated, to resize the text large enough for OCR, and allow the user to easily select it and get an immediate lookup , the whole process would be a whole lot slower and more tortuous than simply doing a pinyin lookup.

Well if we were playing the video then we'd already presumably have it correctly-oriented - it would just be a matter of resizing a box on the screen to indicate where the text would be and our system could do the rest.

Usually, it's a single phrase or character in the subtitle you want to look up ( if you need to look up every character or phrase in almost every subtitle, you'll give up soon enough) . Also, I don't think Pleco is designed to do translation of whole sentences or phrases, such as appear in subtitles, so you'd probably not have it doing automated subtitle translation.

We could send them to Google Translate, I suppose, though it's admittedly not ideal... but if you're watching Chinese movies for learning purposes you're really just going to want those single-word lookups, and I think as long as we can get the OCR to happen automatically we can make that fast enough to be usable.

But basically, I'd rather take the extra time to do something like this than just embed an ordinary video player into our app and call it a day.

 

[mikeo]

We could send them to Google Translate, I suppose, though it's admittedly not ideal... but if you're watching Chinese movies for learning purposes you're really just going to want those single-word lookups, and I think as long as we can get the OCR to happen automatically we can make that fast enough to be usable.

As a solution to translating phrases, or maybe current usage that's not in the dictionaries, GT makes sense, but for just looking up common words or phrases, it wouldn't seem necessary, since you have all these great Pleco lookup tools locally and don't need to pay any latency tax or be restricted to watching movies only where there's an Internet connection.

PS: noticed that Pleco "translate" doesn't return the pinyin of the characters, just English translation....any way to add pinyin?

But basically, I'd rather take the extra time to do something like this than just embed an ordinary video player into our app and call it a day.

+1, totally agree. Still, this whole lookup of video subtitles process is pretty painful, sometimes the nonoptimalism makes me wince.

 

[flameproof]

We've got a lot more in PLC than in ABC - are those still not sufficient? Tuttle is also full of them, and a few of the new dictionaries we're launching this summer have a bunch as well.

I find the sentences in ABC and PLC useful, as in 'better then none', but often also very cryptic.

I think a user created sentence DB could be useful. For example, when you test flashcards (characters) you have plenty of empty space on the screen, there could be a 'hint' function (which is empty). There a user could put in their own hints. I i.e. would just put in a few Chinese sentences only (no Pinyin, no English). Other people may like to have other hints.

 

[character]

I realize this is not a good way to judge dictionary entries, but I'd like a feature which showed me the longest (most words) definition first, and let me cycle through dictionary entries in order of length.

I'd like a flashcard feature which randomly showed me the Simplified or Traditional version of a word/phrase during a test.

 

[keyclick]

I realize this is not a good way to judge dictionary entries, but I'd like a feature which showed me the longest (most words) definition first, and let me cycle through dictionary entries in order of length.

Today I was thinking a similar thing... it would be nice to be able to 'fast forward' to the 4 character+ entries while doing a 字 page lookup. I often search for a 成语 using a specific character, and it can be a whole lotta scrolin' before you get to the 4 character zone. Then, to switch dictionaries, you end up at the top of the list again!

 

[mikelove]

PS: noticed that Pleco "translate" doesn't return the pinyin of the characters, just English translation....any way to add pinyin?

We could, but it's not that accurate, not really even much better than what we could do ourselves offline; I suppose at some point I have to drop my phobia about this and add support for auto-generating confessedly-inaccurate Pinyin since there've been so many requests for it, but I'd like to hold out on that for as long as possible.

I think a user created sentence DB could be useful. For example, when you test flashcards (characters) you have plenty of empty space on the screen, there could be a 'hint' function (which is empty). There a user could put in their own hints. I i.e. would just put in a few Chinese sentences only (no Pinyin, no English). Other people may like to have other hints.

That's something we've been thinking about for a while - the big design question is how we integrate this with the flashcard system; do we add notes / extra examples / etc to entries via the flashcard system and query the flashcard database for them when displaying entries, or do we maintain a separate database of entry notes (tied to some combination of headword / pronunciation or to the dictionary entry id) which exist independently of flashcards? The latter approach might be more flexible, but the former saves time and would probably also be a bit faster in execution since a single query could both find those notes / extra examples and let us provide an indicator to tell you whether or not a flashcard already existed for a particular word (another oft-requested feature).

I realize this is not a good way to judge dictionary entries, but I'd like a feature which showed me the longest (most words) definition first, and let me cycle through dictionary entries in order of length.

That would be very dicey at the moment because we've got two or three different formats in place for encoding dictionary entries in databases - if we literally just compared lengths you might see some seemingly-short ABC entries sorting ahead of seemingly-much-longer 21C ones.

I'd like a flashcard feature which randomly showed me the Simplified or Traditional version of a word/phrase during a test.

That's a neat idea, but how would we handle this in terms of score tracking? Should getting either version wrong hurt the card's overall score, or would we have to break it down separately?

Today I was thinking a similar thing... it would be nice to be able to 'fast forward' to the 4 character+ entries while doing a 字 page lookup. I often search for a 成语 using a specific character, and it can be a whole lotta scrolin' before you get to the 4 character zone. Then, to switch dictionaries, you end up at the top of the list again!

If you tap on the button right next to the dictionary icon to show the (oft-criticized) fast scrollbar, that would let you jump down to the 4-character entries a lot more quickly than you can by manually thumbing through.

 

[character]

That would be very dicey at the moment because we've got two or three different formats in place for encoding dictionary entries in databases - if we literally just compared lengths you might see some seemingly-short ABC entries sorting ahead of seemingly-much-longer 21C ones.

OK. Then I'll add my vote for someone's suggestion of showing a listing of all dictionary entries for a word in a long list.

I'd like a flashcard feature which randomly showed me the Simplified or Traditional version of a word/phrase during a test.

That's a neat idea, but how would we handle this in terms of score tracking? Should getting either version wrong hurt the card's overall score, or would we have to break it down separately?

At first it seemed obvious that getting it wrong should just hurt the overall score, but then I thought about someone consistently getting only the simplified version wrong. I don't have a strong opinion as I've not delved into Pleco's stats, but with spaced repetition I could imagine a scenario where the user doesn't see the version they get wrong often enough to learn it.

How about this as a starting point for the feature: just use the overall score, and randomly show either version? I would think people who are really focused on their stats would also be more likely to focus on one character set at a time.

I guess the alternative for me would be to switch the display settings manually to only show one character set or the other every few passes through the flashcards.

 

[mikelove]

OK. Then I'll add my vote for someone's suggestion of showing a listing of all dictionary entries for a word in a long list.

That actually should be coming, along with search results that show one copy of each word found in any dictionary - it's really only ever been a matter of waiting for mobile processors to be fast enough, and I think they're about there now.

How about this as a starting point for the feature: just use the overall score, and randomly show either version? I would think people who are really focused on their stats would also be more likely to focus on one character set at a time.

That's probably easiest, though another feature that's been in the "maybe" list for a while had been allowing one to mix multiple profiles in a single session - different test types potentially counting towards different profiles; lots of work required to get the statistical stuff right, though.

 

[mikeo]

PS: noticed that Pleco "translate" doesn't return the pinyin of the characters, just English translation....any way to add pinyin?

We could, but it's not that accurate, not really even much better than what we could do ourselves offline; I suppose at some point I have to drop my phobia about this and add support for auto-generating confessedly-inaccurate Pinyin since there've been so many requests for it, but I'd like to hold out on that for as long as possible.

Since the "translate" is starting with knowledge of the selected characters, could you do the pinyin lookup offline and combine it with the English translation from online "translate'?

 

[umibozuuu]

Is there a feature in the Document Reader to scroll like the pages of a book?

Something like, right-to-left finger swipe "turns" to the next "page" and left-to-right finger swipe "turns" to the previous "page"? Outside of those two swipes, no finger action would cause scrolling (in fact there would be no scrolling just page turning..). I really think everything about Pleco right now is already the bee's knees, but I want to get an older person to start using the Document Reader , and I'm sure this page turning would be a winner (even more on iPad..?) versus scrolling, which is just unnatural to her. With page numbering and the creasing of paper that would be perfect :lol: . I can imagine it's much easier to do if only allowed in "lock orientation" mode though , which doesn't bother me much, as the older person I'm thinking of would just be annoyed/perplexed at the screen tilting :mrgreen: .

Feels like a silly request, but that's your fault for having gone beyond what I had imagined and done it all well :wink:

 

[mikelove]

Since the "translate" is starting with knowledge of the selected characters, could you do the pinyin lookup offline and combine it with the English translation from online "translate'?

Sure, but my point is that our version would be inaccurate too - there isn't a way to accurately convert characters to Pinyin at the moment, at least nobody seems to have come up with one. But I suppose if we're going to put up with that anyway it'd be easier to fetch both from Google in the Translate box at least.

Is there a feature in the Document Reader to scroll like the pages of a book?

Something like, right-to-left finger swipe "turns" to the next "page" and left-to-right finger swipe "turns" to the previous "page"? Outside of those two swipes, no finger action would cause scrolling (in fact there would be no scrolling just page turning..). I really think everything about Pleco right now is already the bee's knees, but I want to get an older person to start using the Document Reader , and I'm sure this page turning would be a winner (even more on iPad..?) versus scrolling, which is just unnatural to her. With page numbering and the creasing of paper that would be perfect . I can imagine it's much easier to do if only allowed in "lock orientation" mode though , which doesn't bother me much, as the older person I'm thinking of would just be annoyed/perplexed at the screen tilting .

There isn't at the moment, and basically the reason why is that we think it's extremely important when reading a book in a foreign language that you be able to see an entire sentence / paragraph on the screen at once. It's even more critical in Chinese, where even a single word can potentially span across two different pages.

If we can come up with an elegant interface design that would combine the ability to see the next few lines of text with page-turning then we could certainly take a run at adding that, but we haven't really settled on one that we like yet; showing the next few lines grayed out in the header / footer is the best we've come up with so far, but we're worried that'll make it very easy to lose one's place in the text.

 

[umibozuuu]

If we can come up with an elegant interface design that would combine the ability to see the next few lines of text with page-turning then we could certainly take a run at adding that, but we haven't really settled on one that we like yet; showing the next few lines grayed out in the header / footer is the best we've come up with so far, but we're worried that'll make it very easy to lose one's place in the text.

I was thinking of something trivial like the behaviour you get with only "Page scroll buttons" on in iPod, remapping the page scroll buttons to finger swipes, and blocking any other form of scroll. Solutions of graying header/footer could become confusing: eg. what if I have a really long sentence in a really big font? I didn't anticipate your whole sentence wish..

The "hardware scroll" "(only)" on Windows Mobile (Axim X3) is pretty close to page turning and the popup definition does reach into the unseen next page to complete a word for definition as need arises. If you want the whole sentence you can chose to plod along in the popup def window anyway.
This won't work on iOS because the popup def on iOS doesn't copy the whole sentence. And it looks a lot nicer this way!

Now my following suggestion will seem a lot of hassle for this tiny feature (and it'll be confusing anyway): just throwing it out there for competition with gray header/footers..
Say in "page-turning mode", when you are looking at the popup def of the last character/word on the page/overlapping the end of the page, pressing the right arrow has the effect of switching to a different popup def window, like that in the Windows Mobile reader, that has a copy of the sentence in a little included box. I don't know if it's convenient to allow innumerable presses on the right arrow or block them once end of sentence is reached, but the effect of left arrow could be that "when you come back to a character/word on the page", then switch back to the kind of popup def that is normal for iOS.

 

[mikelove]

The "hardware scroll" "(only)" on Windows Mobile (Axim X3) is pretty close to page turning and the popup definition does reach into the unseen next page to complete a word for definition as need arises. If you want the whole sentence you can chose to plod along in the popup def window anyway.

That was our solution to the same problem on WM (seamless scrolling would have been difficult with fidgety resistive touch screens), but it felt very cumbersome and really didn't let you see the whole sentence anyway (just a few characters at a time) so we dropped it on iOS.

Another idea would be to have an on-screen button to temporary "lock" to page scrolling / disable free-form scrolling but retain the ability to switch back to free-form scrolling with another button press; that way you could use pages most of the time but have the flexibility to leave them to reposition the document when reading a particularly difficult sentence / paragraph. The main design question with that approach would be whether we should stick with a fixed set of page boundaries for the entire document (jumping to one of them when you tap on the "lock" button) or whether we should simply move a page worth of lines back / forward from your current location and not jump around at all when you tap on "lock."

 

[alanmd]

Has anyone ever brought up the idea of a "Universal Search"?
The multi-dictionary aspect of this has been in the pipeline for a while; we've basically been waiting on processor speed, we need to be able to not only search every dictionary instantly (which we already do in a lot of places - pretty much any time you look up anything with the document reader or with OCR you're doing an all-dictionary search) but also merge together instances of the same word; we think it's a lot more useful this way, get a list of words independent of dictionary and tap on one of them to view every dictionary's results for it on a single page.

That sounds like an excellent feature, I pretty much just stick with one of the installed dictionaries most of the time at present.

 

[johnh113]

Dear Mike,

Some minor comments:

1. I'd like to be able to go back to the input screen and add a tone on the first character. For instance, let's say I input nihao in pinyin with no tone marks. Then I decide I want to be more specific and use your keyboard to add the 3rd tone marker after ni. It isn't allowed. The 3rd tone marker goes at the end of the word. I can't find a way to insert the tone marker into the middle of the pinyin.

2. I can't remember what I was looking up or even whether it was hanzi, pinyin or English, but after selecting the text by pressing and holding to get the copy/search icon, I was at the end of the line of text, and I couldn't figure out any way to drag the little blue selection dot so that it would wrap around and include the text at the beginning of the next line.

3. In flashcards, when I tap on a character to select it and the pop-up window pops up, I get 7 icons along the bottom (because of the add to "copy to pasteboard" icon that was added in the last release). 7 seems a bit much, but the real problem for me is that I use the audio all the time and hit the audio icon. I often accidentally hit the + icon or the enlarge text icon instead because they are squished together so much. Both of these icons require work to undo. The + has now added a word that I didn't want so I have substantial work to the organize cards page and delete the flashcard I didn't want. The enlarge icon isn't as bad as I only have to touch the screen to get back where I want, i.e. just one extra keystroke. If copy to pasteboard were next to the speaker it wouldn't be so bad because I don't care about undoing that and I can just directly hit the sound icon again. Or the active area around the speaker icon could be larger because accidentally hitting the speaker icon doesn't do anything that needs to be undone and is a harmless mistake. I like the way the flashcards screen, at the top, where there are 7 icons, alternates buttons with display only information so that even though there are 7 separate fields, I don't hit a wrong button. You could probably even enlarge the active area around the buttons to make them even easier to hit, but they're fine for me right now.

(I had held off on UI comments because you had hired an outside company to help design the UI (It didn't surprise me that they couldn't help. I was skeptical from the beginning as the program is very complex and would require someone intimately familiar with how people use the product.))

John

 

[mikelove]

1. I'd like to be able to go back to the input screen and add a tone on the first character. For instance, let's say I input nihao in pinyin with no tone marks. Then I decide I want to be more specific and use your keyboard to add the 3rd tone marker after ni. It isn't allowed. The 3rd tone marker goes at the end of the word. I can't find a way to insert the tone marker into the middle of the pinyin.

That'll be fixed as soon as we drop support for versions of iOS prior to iOS 4; the issue now is that we actually have very little control over what happens in that text input field since Apple didn't officially add support for custom input methods before OS 4.

2. I can't remember what I was looking up or even whether it was hanzi, pinyin or English, but after selecting the text by pressing and holding to get the copy/search icon, I was at the end of the line of text, and I couldn't figure out any way to drag the little blue selection dot so that it would wrap around and include the text at the beginning of the next line.

Couldn't you just drag it downwards, or were you at the bottom of the page?

3. In flashcards, when I tap on a character to select it and the pop-up window pops up, I get 7 icons along the bottom (because of the add to "copy to pasteboard" icon that was added in the last release). 7 seems a bit much, but the real problem for me is that I use the audio all the time and hit the audio icon. I often accidentally hit the + icon or the enlarge text icon instead because they are squished together so much. Both of these icons require work to undo. The + has now added a word that I didn't want so I have substantial work to the organize cards page and delete the flashcard I didn't want.

Undo for flashcards is a heavily-requested feature, so that might help a little bit on this, though really we just need a way to squeeze more functionality in with fewer icons - probably either by letting you disable some less-frequently-used ones (like the zoom button) or by burying a few of them in a sub-menu (tap on a button and then another bar pops up with more options).

(I had held off on UI comments because you had hired an outside company to help design the UI (It didn't surprise me that they couldn't help. I was skeptical from the beginning as the program is very complex and would require someone intimately familiar with how people use the product.))

We were hoping they could at least come up with a nicer color scheme / UI skin / etc for us, but they didn't even manage to do that, and their first prototype for a new app icon was a considerable step down from what we have now so we didn't end up hiring them to redesign our other icons either. We would still love to find a good icon designer at least - skins and color schemes are one thing, but the fact that all of our current icons were designed by programmers is obvious and a little embarrassing

 

[alanmd]

1. I'd like to be able to go back to the input screen and add a tone on the first character. For instance, let's say I input nihao in pinyin with no tone marks. Then I decide I want to be more specific and use your keyboard to add the 3rd tone marker after ni. It isn't allowed. The 3rd tone marker goes at the end of the word. I can't find a way to insert the tone marker into the middle of the pinyin.

Not a fix, but as a workaround you can disable the feature that changes the last tone when you click the tone buttons. This makes changing the tone on the last character a bit more awkward (you have to press backspace to get rid of the previous tone number), but makes changing the tones on previous characters a lot easier.