Flash Card Program For Testing Vocabulary?

[Frederic Bergeron]

I think you will find porting to QT4 a mixed bag. Many functions changed name, some disappeared.
For some classes you will have to go to a lot of trouble to find a work around and others will be easy.
I recommend avoiding the QT3 support libraries, you won't learn QT4 that way.

Good luck!

Thanks!  The documentation seems to contain very valuable information about how to port application from 2.x to 3.x and 3.x to 4.x.  I will read that beforehand to avoid unpleasant surprises.  I will follow your advice as I don't care about Qt3.  I don't intend to use qt3to4 tool either.  I will do the conversion manually in order to see the differences.  Hopefully, it won't be too hard to do.

[kurochka]

Yokatta!!!

The search function, though basic, is great news!  I constantly find myself confused by similarly sounding words and need to find what that other word was.  This would be helpful.  Show alt in the list option is also useful.  Now I can choose to see romaji.

It's unfortunate that KanjiCafe is so inflexible.  I could suggest showing the copyright notice every time the quiz containing these images is started that should be often enough; however, it seems they are set in their ways.   In any case, they distribute the images for free to end-users so I will use them on my own.  Any help will be appreciated though.  I think distributing glossaries with links without images should work well.  There are some precedents with similar approaches (e.g., provision for use of some closed drivers in Linux, Microsoft fonts also in Linux, etc.)

[Frederic Bergeron]

Yokatta!!!

The search function, though basic, is great news!  I constantly find myself confused by similarly sounding words and need to find what that other word was.  This would be helpful.  Show alt in the list option is also useful.  Now I can choose to see romaji.

I'm glad you like the new version.  Now that the Search Tool is there, I use it quite often too.

It's unfortunate that KanjiCafe is so inflexible.  I could suggest showing the copyright notice every time the quiz containing these images is started that should be often enough; however, it seems they are set in their ways.

Indeed :-(

In any case, they distribute the images for free to end-users so I will use them on my own.  Any help will be appreciated though.  I think distributing glossaries with links without images should work well.  There are some precedents with similar approaches (e.g., provision for use of some closed drivers in Linux, Microsoft fonts also in Linux, etc.)

I agree with you.  Concerning copyrights, I think it would be legal as there would be no image copies involved.  However, it would still be a violation of the licence agreement.  I don't think that there are pertinent legal consequences against that though, are there?  I would not be listed as a good licensee but eh!  If that's the price to pay...  Anyway, I will think about it and ask advice to as many people as possible.  

Meanwhile, if you want to use the images for your own glossaries, it's rather tedious but it's possible.  I use this online tool to get the UTF-8 character number associated to a specific kanji so that I can find the corresponding image file.  Some kanjis have no images either because no volunteers have done them yet or because they are not jouyou kanjis.

In time, I would like to develop a plug-in that would alleviate this task.  It would be an independant component external to toMOTko that you would install separately.  toMOTko would detect if it's there or not.  If so, it would be possible to invoke a special interface where you will enter the kanji and the corresponding image would be automatically fetched (if it's found).  That will have to wait though until I complete the port to Qt4.

By the way, so far, it works fine.  I can already see (partially) the Glossary Manager on Windows and Linux.

[img]http://www.fbergeron.com/zaurus/tomotko/toMOTko_qt-4_win_demo-1.png\" border=\"0\" class=\"linked-image\" /]

Updated : Here is how it looks (2008/11/27)

[img]http://www.fbergeron.com/zaurus/tomotko/toMOTko_qt-4_win_demo-2.png\" border=\"0\" class=\"linked-image\" /]

[kurochka]

it would still be a violation of the licence agreement.  I don't think that there are pertinent legal consequences against that though, are there?  I would not be listed as a good licensee but eh!

License agreement is just that -- an agreement or contract of two or more parties.  In order to use the images, you have to agree to that agreement.  If you do not use the images in your application you do not enter into the agreement.  Therefore, you cannot violate it by doing to your application whatever you want to.

Copyright is different from license agreements because it does not require your agreement/consent/acceptance.  Again you do not violate copyright if you do not copy, distribute, etc. the copyrighted works themselves.  There is an argument that by creating a program/function that can have no other legitimate use other than to permit copyright infringement by third parties, you facilitate/infringe on that copyright (e.g., some of the claims in the lawsuits against Kazaa).  However, that's an extremely hard argument to prove.

[kurochka]

I think there might be a bug/problem with 0.11.  I uninstalled the 0.10, installed 0.10.  I didn't notice any recompilation of the study lists.  When I tried to start a quiz, there was a request whether I wanted to resume the old one.  Whether I said yes or no, the application would shut down.

Then I went to the Preferences section and did something to Revealing sequence (I think) and the application started working.  Maybe default settings after the upgrade made it touchy.

Also, could you give some explanation, examples of the new revealing sequence dialog.  Should I make it something like a+c+b?  Is just a enough?  If just a is enough, what will be revealed after a?

Just saw a pic on http://tomotko.sourceforge.net/img/screens...erencesQuiz.png

I think + means "together with" and > means "before/earlier".

Could you put in some default sequences that we can choose from?  When I just installed the application the sequence field was empty.

[Frederic Bergeron]

After reading all your message, I think the shutdown occurred because you had no Revealing Sequences defined.  Indeed, toMOTko could not proceed to resume the quiz because it didn't know where to start (which fields to reveal first).  That's odd because there should have been 3 sequences defined by default after installing 0.11.0 :

a > c+e > b+d (this one should be checked)
a+e > c > b+d
c+e > b > a+d

Also, toMOTko would not allow you to close the Preferences dialog if you delete all the Revealing Sequences.  So you're forced to define at least one.

If you don't mind, could you compress and send me (by email) the directory /home/zaurus/.toMOTko-before0.11.0 if it exists.  As the name implies, it's a copy of your data before you install the latest version.  This way, maybe I can reproduce and fix the bug.

About the Revealing Sequence, I think you understand pretty well.  Just in case, here is an extract from the Preferences page of the Online Help (found on the website or directly in the application if you click the question mark in the upper right corner next to close button) :

A revealing sequence determines which fields will be shown by the Quiz when asking a question and in which order they will be revealed. A revealing sequence is expressed as a character string similar to a mathematical expression. Each letter corresponds to a specific field of the Quiz as shown in the picture on the left. Between each letter, either a junction (+) operator or an iteration (>) operator is used.

For example, let's take a look at the following revealing sequence : a+e > c > b+d.

a+e means that the Quiz, when asking a question, will start showing the source language's term field (a) and the image field (e) (if there is one defined for the current term). Then, if the Eye button is clicked, the target language's term will be revealed ©. If the Eye button is clicked again, the target language's Alt./Phon. field (b) and Comment field (d) will be revealed.

When all fields are revealed, if Eye button is clicked another time, the Comment field will be expanded. If the Eye button is clicked again, the Comment field will be minimized.

If you have a Revealing Sequence that is only a, only the field (a) will be shown by the Eye button.  The other fields can be shown if you click on them individually.

I hope that clarifies.  I know it's a little bit more complicated than before but it's also more flexible.  If I had kept the previous strategy (using simple checkboxes), it was becoming too cluttered because of the additional combinations that the image field brings.

[koan]

It's unfortunate that KanjiCafe is so inflexible.

There is a better solution, unfortunately I have not had enough time to pursue it yet.

Wikimedia has a project to generate diagrams for stroke order for CJK characters: http://commons.wikimedia.org/wiki/Commons:...e_order_project

It's already used in Japanese wiktionary, for most kanji you can find a nice diagram showing the strokes. They haven't drawn all the diagrams but I am fairly confident they will finish the project.

I noticed this because I started looking into converting Japanese wiktionary for zbedic. Reading through wiktionary dumps is quite challenging; some people have already written parsers; I found quite an advanced one and downloaded the source but have not had the time to  investigate it (it's not a simple matter to just run it and generate output).

What needs to be done is
 

• Write a parser to go through wiktionary and find linked kanji
• Script to download kanji diagrams from wikimedia (only links are in wiktionary)
• Script to cut diagrams into separate frames, resize them and make an animgif (toMOTko can handle GIF89 right ?)
• Script to generate flashcard file with pictures.

Wikimedia project has already generated >300 animgifs but they have about 1000 kanji in total. Most of these steps would be required if you use kanjicafe's images.

If you have more time than me and want to do this work, then please tell me so I don't duplicate the effort. However, I can't give an estimate of how long it will take me. Maybe a very long time!

[Frederic Bergeron]

License agreement is just that -- an agreement or contract of two or more parties.  In order to use the images, you have to agree to that agreement.  If you do not use the images in your application you do not enter into the agreement.  Therefore, you cannot violate it by doing to your application whatever you want to.

Copyright is different from license agreements because it does not require your agreement/consent/acceptance.  Again you do not violate copyright if you do not copy, distribute, etc. the copyrighted works themselves.  There is an argument that by creating a program/function that can have no other legitimate use other than to permit copyright infringement by third parties, you facilitate/infringe on that copyright (e.g., some of the claims in the lawsuits against Kazaa).  However, that's an extremely hard argument to prove.

Interesting.  As the application is not aware if an image associated to a term comes from KanjiCafe or not, I imagine that I'm not bound to the license agreement.

However, in the case I develop a plug-in facilitating the association between toMOTko and KanjiCafe's images, I could hardly deny that.  Or could I?  It's arguable as, once again, the plug-in doesn't need to be aware of KanjiCafe in reality.  It just needs to convert a kanji to its UTF-8 character code to find a corresponding file that could be KanjiCafe's or not.  As a matter of fact, it could even be a CJK stroke order project's file (mentioned by Koan's last post).  

There is a better solution, unfortunately I have not had enough time to pursue it yet.

Wikimedia has a project to generate diagrams for stroke order for CJK characters: http://commons.wikimedia.org/wiki/Commons:...e_order_project

It's already used in Japanese wiktionary, for most kanji you can find a nice diagram showing the strokes. They haven't drawn all the diagrams but I am fairly confident they will finish the project.

Thanks Koan for this link.  It looks full of potential.

• Write a parser to go through wiktionary and find linked kanji
• Script to download kanji diagrams from wikimedia (only links are in wiktionary)
• Script to cut diagrams into separate frames, resize them and make an animgif (toMOTko can handle GIF89 right ?)
• Script to generate flashcard file with pictures.

Why do I need steps 1 and 2?  Can't I download the files directly from the website using wget?  Maybe they are not up to date?  

EDIT: The version of wget I'm using (1.10.2) doesn't seem to digest the asian characters in the urls so the images cannot be downloaded :-(  Maybe an alternative tool could achieve that though.

For step 3, Yes, toMOTko can handle GIF89.  For this job, ImageMagick is THE tool as mentioned in this page :

http://wiki.flux-cms.org/display/BLOG/Resi...ith+ImageMagick

I tested the procedure and it works well.  Result is better when reducing an image but is still very good (a bit blurry) for making an image larger.  So writing a script to process all the images should not be too hard.  Anyway, as the original size of the images (those found on the website) is 100x100, I think it's already good enough and doesn't need to be resized for toMOTko.

For step 4, it depends...  I don't think we necessarily want to "generate" a flashcard file.  I think it would be better (more generic) to allow a user to easily associate an animated image to a term whenever he wants (that's the idea of the plug-in already mentioned.)    

I think toMOTko could use both sets of animated images.

[koan]

Why do I need steps 1 and 2?  Can't I download the files directly from the website using wget?  Maybe they are not up to date?

Flashcards that only show the stroke order are not useful to me. I want to be able to look up  a kanji, it's meaning and how to write it.

I don't have a big list of kanji so if I don't parse wiktionary I don't know what files to download.

EDIT: The version of wget I'm using (1.10.2) doesn't seem to digest the asian characters in the urls so the images cannot be downloaded :-(  Maybe an alternative tool could achieve that though.

Have you tried representing those characters in escaped UTF8 ? http://www.w3.org/TR/html40/appendix/notes...non-ascii-chars

For step 4, it depends...  I don't think we necessarily want to "generate" a flashcard file.  I think it would be better (more generic) to allow a user to easily associate an animated image to a term whenever he wants (that's the idea of the plug-in already mentioned.)

As I said, my plan is to make a flashcard file with kanji, reading and stroke order. I'm not interested in a plugin to select kanji images to attach to a term.

[kurochka]

Frederic,

at the end of your post here (a way back):

https://www.oesf.org/forum/index.php?showto...mp;#entry170486

you said that you could convert lists listed at this site:

http://www.kahome.co.uk/japa/tango.pl?mo=0...fm=16&to=30

Would you look at this again and see if this is something that you could do?

[Frederic Bergeron]

Sure, I will take a look at it and try to write a conversion script.

[kurochka]

Thanks.

I think that you could delete romaji; it is a bad crutch.

[Frederic Bergeron]

Thanks.

I think that you could delete romaji; it is a bad crutch.

I agree with you that romaji should not be used.  I prefer to keep it though for the sake of other people who think differently.

Here are the Japanese for Busy People and Japanese Now! glossaries.  Others will follow.

[Frederic Bergeron]

Here is the last glossary: Japanese for College Students: Basic.

For the Nihongo Shoho glossaries, it will take more time.  As stated in a previous message (a few months ago), I had already began to convert them from a different source (which provided the words in English and Spanish) but it is a tedious and time-consuming task that cannot be fully automated.  I will try to put some time on it every week to get over it.

The other glossaries listed on the kahome website are redundant.

[kurochka]

Are you talking about our discussion in Nov. 2007 (below link) or have you found other sources?

https://www.oesf.org/forum/index.php?showto...mp;#entry170720