Android Application using KanjiVG

[atomusk]

Hi,

I am the developer of an Android application aimed for people learning
Japanese.
In the first version 4 month ago, it only helped for kana revision,
but 2 month ago I posted on the Android Market an update that use data
from both KanjiDict and KanjiVG to have data on the 2230Kanji of JLPT
4-1.

I know I should have made the announcement long ago, but I never quite
found the courage to do it ;)

Today I have posted a new version again that uses again the data from
KanjiVG to add a drawing recognition for a Kana test (and soon for
Kanji Test).

I already contacted some of you for other purpose (Ben Bullock, your
Kana to Romaji algorithm works like a charm - and thanks to Alex to
have added me in the projects using KanjiVG).

I want to give my thanks this community that enabled me to have an app
that have 23 000 DL, and have been rated 363 times with an average of
4.68 stars.

Unfortunately for now my code is not open source principally because
my code is in no shape to be accessible to the public (Obenkyo is my
first Java project so it's more hack than code), but if someone want
some help implementing KanjiVG in his Android App, I will be more than
happy to share my code.

Regards

Atomusk

[Alexandre Courbot]

Hi,

> I am the developer of an Android application aimed for people learning
> Japanese.
> In the first version 4 month ago, it only helped for kana revision,
> but 2 month ago I posted on the Android Market an update that use data
> from both KanjiDict and KanjiVG to have data on the 2230Kanji of JLPT
> 4-1.

Thanks for publicizing your project. It is always heartwarming to know
that KanjiVG is under use.

> Unfortunately for now my code is not open source principally because
> my code is in no shape to be accessible to the public (Obenkyo is my
> first Java project so it's more hack than code), but if someone want
> some help implementing KanjiVG in his Android App, I will be more than
> happy to share my code.

There is nothing to be shy about - if anything, exposing your code
will allow bugs to be fixed more quickly and will help you becoming a
better programmer. I hope to be able to browse through your git repo
soon. ;)

Alex.

[Ivan Meredith]

Don't be worried about the state of your code. ALL developers have code that "should" be "better" hehe.

You might get people submitting patches or recommending better/different ways to code it :) 

I personally am interested in the drawing recognition part. (I have no idea how to do this and haven't been able to find a good example).

Ivan

p.s This is my first post too :) I've been following the ML for quite awhile now. I have some plans that might involve the data if I every get around to it heh.

--
You received this message because you are subscribed to the "KanjiVG" group.
For options and unsubscribing, visit this group at
http://groups.google.com/group/kanjivg

[atomusk]

On 31 jan, 10:59, Ivan Meredith <i...@ivan.net.nz> wrote:
> Don't be worried about the state of your code. ALL developers have code that
> "should" be "better" hehe.
>
> You might get people submitting patches or recommending better/different
> ways to code it :)
>
> I personally am interested in the drawing recognition part. (I have no idea
> how to do this and haven't been able to find a good example).
>
> Ivan
>
> p.s This is my first post too :) I've been following the ML for quite awhile
> now. I have some plans that might involve the data if I every get around to
> it heh.

Well ... it's worst the a matter of I would like it "better" ... ;-)

Concerning the drawing recognition, I tried to explain the differents
methods I used to "differenciate" a kana for another in my blog :

http://atomusk.free.fr/?p=73

Since I am in a "test" scenario, I just have to check if the kana that
was drawn is close to the Kana.
So, I can check stroke number, length, heigth, width, and angle of
each stroke. and if it is close enougth, (with quite wide margins), I
consider it good.

Do you have plans for android ? Or just by curiosity ?


Atomusk

[Ivan Meredith]

Actaully I just bought an iPhone :) I used to have an android, but apple stuff 'just works', for all its other failings.

I'm more interested in web based applications.

--

[Alexandre Courbot]

On Sat, Feb 16, 2013 at 12:21 PM, Andree Surya <andree...@gmail.com> wrote:
> Hi Atomusk,
>
> I would like to incorporate KanjiVG for my upcoming Android app.
> However, I believe it's going to be a non-trivial task since Android SDK
> doesn't provide native support for SVG rendering.
>
> I have seen your app - Obenkyo. Your Kanji stroke order animation is really
> cool!
> I'm interested in knowing your approach to parse KanjiVG files. Did you
> parse KanjiVG by yourself, or did you use SVG parsing library?

Parsing SVG (especially the subset used in KanjiVG) is not hard,
really. For Android you just need to translate the SVG commands into
the appropriate calls to Path methods to get the shapes. Doing the
animation itself is another business, I used a DashPathEffect and an
ObjectAnimator to obtain something smooth and efficient.

What I describe here is actually done by the (not public yet) Android
version of Tagaini. It already uses the Tagaini databases so a lot of
the work does not need to be duplicated. I don't know how advanced
your project is, but if this does not conflict this your goal may I
suggest you consider contributing to this project instead of writing
another jmdict/kanjidic/KanjiVG Android app. There are many different
ones available already, none of which is particularly impressive. I'd
really like, at some point, to federate some of these and make the
Android-Japanese-app-to-rule-them-all. :P

Alex.

[Andree Surya]

Hi Alex,

Thanks a lot for the direction. I'm currently studying the SVG files and hopefully can start some drawing soon.

I'm currently working on a yet another Kanji flashcard app called Kanji Renshuu. 

I thought it would be useful to incorporate Kanji stroke order information in my app.

Cheers,

Andree

[Alexandre Courbot]

On Tue, Feb 19, 2013 at 3:17 AM, Andree Surya <andree...@gmail.com> wrote:
> I'm currently working on a yet another Kanji flashcard app called Kanji
> Renshuu.
> I thought it would be useful to incorporate Kanji stroke order information
> in my app.

Nice app - is it open-source by any chance?

Alex.

[Andree Surya]

Kanji Renshuu is closed-source, but the idea of open-sourcing the app (parts or entirety) is something that I'd like to consider.

Currently we already have one great open source flashcard app in Android platform - AnkiDroid. It has a good SRS implementation, but a little bit bloated because it aims to be generic (not just for learning Kanji) and has many features I don't need. Kanji Renshuu, on the other hand, aims to be simple, comfortable, and just enough for my learning needs. Probably if I could extract the generic flashcard component out of Kanji Renshuu and open source it, others can develop something similar for other purposes.

- Andree