Making flashcards fun and challenging again: Character Recognition Workout

[character]

Hey, I don’t post much, though I’ve been on the forum for years. You might know my Graded Chinese Readers post. Anyways, I built an app which I think has an interesting take on learning to recognize characters. In Chinese class I saw students struggling to recognize characters if they looked even a little bit different from what they were used to. While much of the app is geared towards beginners, there are some unique features which may appeal to more experienced learners as well.

App Store link

 

<Sales pitch>

Standard Chinese flashcards are great…at first. But seeing a character presented in the exact same way, over and over again, would be perfect if that was how you encountered Chinese in the real world. Chinese in the real world is quite different; the same character can appear in a wide variety of shapes and sizes. I had the idea to make a flashcard program to help you cope with recognizing Chinese characters in the real world: Character Recognition Workout - Mandarin Chinese Flashcards. By transforming characters so they don’t always look the same, the program gives your character recognition skills a workout, so recognizing characters in the wild is easier. It is like how you work out with weights or on a treadmill to be better equipped to handle everyday challenges. You can also leave characters untransformed, for the classic boring old flashcard experience. It’s up to you. Transformations are things such as warping, cropping, rotating, and masking.

While this feature alone would make it worth getting the app, Character Recognition Workout - Mandarin Chinese Flashcards also has an on device two person mode where you and someone else can compete head to head recognizing characters. From personal experience, beating a classmate (or even a native Mandarin speaker) at recognizing characters is very motivational.

The app comes with:

• Over 3,500 flashcards so you can get started immediately
• The popular CC-CEDICT Chinese-English dictionary for adding your own flashcards
• Multiple-choice or review modes for one-person flashcards
• Apple watch review mode
• Two-person single character multiple-choice mode
• Sharing options to save or share character images
• Extensive built-in help
• An experimental/unsupported way to import flashcards sets

</Sales pitch>

 

I expect a typical user would run through their flashcards on Normal once or twice, then switch to more challenging modes for future reviews.

Requires iOS 8.2 or later. If sales indicate it would be worthwhile, I’m open to making an Android version. Intro price is $0.99 in the US (and approximately that in other countries). The discount for educational institutions is enabled for volume purchases.

Interested in conducting an academic study of Character Recognition Workout - Mandarin Chinese Flashcards vs. normal Flashcards? Contact me for suggestions.

Learn more here: http://firstpersoncoder.com/

[vellocet]

Anki has been knocking around this idea, but hasn't really made much progress towards finishing it. This phenomenon has a name, but I forget what it's called.  "I want to train at recognizing a character in any font, not train at recognizing a character in a particular font."  I've noticed it too.

 

You actually stretch the font around and distort it?  I think just substituting another font would be cooler.  I see lots of different ones in advertising, or in venue names.  Or restaurant menu fonts.  You could have the handwritten cursive character in the corner, just to help you along.  I don't know how effective drilling distorted fonts would be. 

[Demonic_Duck]

I don't know how effective drilling distorted fonts would be.

 

Very effective, if your goal is to be able to deal with 验证码 when signing up for websites!

[character]

 

You actually stretch the font around and distort it? [...]  I don't know how effective drilling distorted fonts would be.

The app's default is no transformation (Normal).  The other settings (besides Random) are Crop, Mask, Angle, Fade (displays the character only slightly darker than the background) and Warp.  Only Warp transforms the shape of the character, and I've spent hours picking transforms and parameters to make sure the character is still readable.  The idea is to keep flashcard sessions from being boring exact pattern matching exercises.  By altering the character a bit, it helps you learn to recognize the "essense" of the character, not the exact image of the character.

 

 

I think just substituting another font would be cooler.  I see lots of different ones in advertising, or in venue names.  Or restaurant menu fonts.  You could have the handwritten cursive character in the corner, just to help you along.

 

I think that would be a good additional option.  Based on what other Chinese learning programs charge for each additional font, I didn't think an app with several fonts built in would be seen as affordable by the audience I was trying to reach.  I feel with my app you get most/all of the benefits of different fonts without the cost.

[Demonic_Duck]

Surely there are Chinese fonts out there which are free for commercial use?

 

I disagree that warping the characters is comparable to different fonts. Different fonts will do things that warping characters won't, most notably variant character forms.

[character]

 

I disagree that warping the characters is comparable to different fonts.

I guess we'll have to agree to disagree if you're saying they aren't at all comparable.  Again, the app is aimed mostly at beginners who are trying to get comfortable with characters, not someone like you fine-tuning their Chinese.  I had hesitated to post here about my app due to the negative reaction on the forum that sometimes greets new ideas in Chinese learning.  It's my contention presenting the characters in different ways makes it easier to recognize them in the real world than standard flashcards, and I'm looking to interest academics in conducting studies to see if that is the case.

[vellocet]

Free fonts:

 

https://code.google.com/p/wangfonts/

GPL

 

http://www.wazu.jp/gallery/Fonts_ChineseSimplified.html

[Demonic_Duck]

I had hesitated to post here about my app due to the negative reaction on the forum that sometimes greets new ideas in Chinese learning.

 

My intention is to give criticism that's constructive, not overly negative. Sorry if my tone isn't conveying that intention very well. If you look at my comments in e.g. the GoVoluble thread, you'll see that I dish out lots of criticism even for new products which I absolutely love, in the hope that the feedback will be useful to help the creator of the product improve it even further (I loved GoVoluble so much I linked to it in my signature and will extol its virtues to anyone who'll listen).

 

What I'm trying to say is that I think warping and different fonts are 两回事, and that including support for different fonts (assuming it's viable, e.g. by using the Google fonts linked to in #7) would increase your app's usefulness and widen its appeal.

[character]

 

Free fonts:

 

https://code.google.com/p/wangfonts/

GPL

 

http://www.wazu.jp/g...Simplified.html

Thanks, but those seem either GPL'd (not permitted on the App Store by Apple for use in an app), covered by Adobe, etc. licensing restrictions preventing their use in an app, not findable when I try to download them, etc.  It's my understanding Apple's adding new Chinese fonts in iOS 9 and I had already planned to add them.

 

I do encourage people to use my app for a few weeks before forming an opinion of this one optional feature.  I had created a proof of concept of the app using this feature early on and tested it with my Chinese class and folks at the local Chinese Meetup group.  Only because the feedback was good did I proceed to make the full version of the app I've released. 

[imron]

https://code.google.com/p/wangfonts/

These are also very limited in the amount of characters they cover (very few if any simplified if I recall correctly), which makes them not practical for use in any sort of app like this.

 

 

 

Surely there are Chinese fonts out there which are free for commercial use?

I went through a pretty big hunt for fonts when making Hanzi Grids, and at the time there were only 2 fonts with decent character coverage that could be used and distributed without any restrictions - a Songti and a Kaiti font by Arphic.  Recently, Google has also released their Noto font (a Heiti font), bringing the grand total to 3 - (I'd be happy to hear about more if you can find them.  Edit: forgot, there is also Hannom, bringing the total to 4, but it's also Songti, so really only 3 different styles)

 

Most of the fonts listed in the other link are free to download for personal use (and even then half the links don't work), but definitely not free to distribute in an app.

 

Anyway, the current state of things is that you can forget about any sort of fancy Chinese fonts unless you want to pay for licensing, which will likely set you back several hundred to several thousand US$ a year.

[imron]

I do encourage people to use my app for a few weeks before forming an opinion of this one optional feature

I couldn't seem to find any screenshots of this feature on your website or the iTunes page.  Do you have some examples you could show?

[arrow]

I don't think programmatic transformation is what you called "wild" or "real word" chinese characters.

I suggest you offer a system to allow user to upload their own handwrittings and pass it through some OCR engine to verify.

Then randomly chose one from database for other users to challenge.

You might store the writing in vector to reduce data size.

[character]

 

I couldn't seem to find any screenshots of this feature on your website or the iTunes page.  Do you have some examples you could show?

Thanks, yes, I do need to add more about it.  Been struggling with how to market it in a non-confusing way (and finding time to do so due to spending time finding a new contract). 

 

 

I suggest you offer a system to allow user to upload their own handwrittings and pass it through some OCR engine to verify.

Then randomly chose one from database for other users to challenge.

You might store the writing in vector to reduce data size.

That is a fascinating idea; I'm afraid that kind of complexity would not be a good fit for my app.  I encourage you to pursue creating such an app if that functionality is something you want to exist in the world.

[character]

Here's a look at the optional transformations: http://firstpersoncoder.com/ (about a third of the way down the page).