Home > Development, Plans > What takes so long? :)

What takes so long? :)

I guess this kind of thought often appears in the minds of many users, who are not programmers themselves and are waiting for a new version to pop out of nowhere. In order to be able to release the final v0.57 of the program, I have set some difficult and some not so difficult goals for myself. A single new feature sometimes requires solving a hideous but hidden problem that wasn’t obvious at first. Let’s take for example a seemingly simple addition to the long-term study list.

I have filled a few notebooks (made of paper, not plastic) while studying kanji, because I have decided that I not just want to know how to read them, but also how to write them. Because environmentally friendly (also called green, no idea why) things are so popular these days, I thought that letting students practice inside the program without the need to waste valuable paper, ink and money is a good idea. I wouldn’t have thought about this in the first place, if I didn’t already have a good “hand”writing recognition engine built in (made also by me of course). Adding a single button next to the input box and popping up the recognizer window was cake. The problems surfaced soon afterward.

Take for example the word 日々. Everyone studying Japanese should know what 々 is. It is a repetition symbol for kanji, but it is not kanji itself. Thus, the handwriting recognition engine can’t recognize it as it was made solely for kanji. Why was it made like that and why it couldn’t handle anything else is a lot more complicated than to explain it here, so don’t ask. I didn’t want to force anyone to go and hunt for the 々 character elsewhere to copy and paste it into the input box, so I either had to add a fake kanji, which would make everything more painful later, or modify the whole recognition engine to be able to work with other characters too. So I have decided to do the latter. And if I’m doing it anyway, I will have to include all kana characters too, because typing a word with the keyboard and drawing with the mouse simultaneously is very difficult when you should type and draw 60 of them one after the other.

I can tell that modifying the engine wasn’t easy. It wasn’t too difficult either, just took a lot of time.

I wouldn’t have posted about this all if the problems were to stop here. Unfortunately what comes after this will be even more tiresome. I will have to add all the hiragana and katakana characters, and also all the kurikaeshi (kanji/kana repetition) symbols one by one to the stroke order database. And if my plans become reality and zkanji will handle words that have all kind of symbols and Latin characters in their written form I will probably have to create the stroke order drawing of all the full-width Latin characters and numbers too.

I only want to finish the kana and kurikaeshi symbols till the next release, but it will still take some time.

The compulsory screenshot:

Most users will probably never find this window in any of their menus nor will accidentally open it. Unless you add a [Debug] section in the zkanji ini file and a debug=1 value under it, the stroke order diagram editor won’t show up. There is currently no way to share single diagrams among users, so even if someone wanted to help me by creating the remaining 1200 S.O.D. (5100 is already done!) plus all the kana, they couldn’t send them to me. (Well, they could, and I could modify the program to get the required data from the received file, but that’s more work than how much I would save with it. (Probably not, but I don’t need the remaining 1200 S.O.D. right now and most people don’t either.))

  1. December 17, 2010 at 9:41 pm

    I got zkanji recognize a lot of hiragana now, but I got stuck at り. Whatever I try, it won’t work. I had to make several changes to the recognition algorithm, and now I’m afraid that something might break if anything is changed. So I’m experimenting by editing the SOD. There is nothing much else I can think of right now.

    UPDATE: the second stroke was incidentally written in the wrong direction. Now it works fine and I’m going to finish the whole hiragana table by the end of the day!

  2. Puchatek
    December 22, 2010 at 7:57 pm

    I think there’s a slight bug with deflections search. If you type かえれる zkanji suggests (along with correct deflections) a “potential, potential of 買う”. Well, I’m not an expert in Japanese (まだ :P), but that sounds wrong. I use version 0.537.

  3. December 22, 2010 at 9:00 pm

    Yes I know about it… できれる, which is also recognized by zkanji is incorrect too, though できる appears to be a plain verb. But as the potential form of する it cannot be used. There are several such cases when zkanji does not recognize an incorrect combination and I will have to add a check for them all separately. I just can’t say when.
    If you help finding other incorrect forms that zkanji recognizes, it will make me fix this faster. 🙂

  4. Puchatek
    December 22, 2010 at 11:33 pm

    Will let you know should I stumble upon anything else.

  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: