Dictionary Applications

[koan]

Recently I have been working on ZBedic because I wanted to update it and port it to Qt 5 so I can use it on my desktop. Long story short, porting it was too much hassle so I started a clean implementation. That project is called Frasmodic.

I think it won't be possible to build a Qt 5 version on Zaurus so the best thing to do is to try to update ZBedic. For example, libbedic is statically linked. I can make some improvements to it then build a dynamically linked ZBedic. However, the question is who would use it ? I don't see so many people posting on here saying they are using Sharp ROM and the majority of users seem to use zalarm without GUI (is that correct ?)

I also noticed from web history that someone generated a Japanese-English dictionary from EDICT where they mixed kana and kanji keywords together in the index. I have been generating similar dictionary files myself but I separated the kana and kanji versions.  Having a combined index seems very sensible, I think I can adapt my scripts so that is also in my plans.

Some of my other dictionary files can be improved so I will try to get onto that task in the future. I am also curious about converting FreeDict dictionaries to bedic format, or supporting that format directly.

Finally, I am playing with the idea of porting Kanji Nirvana to desktop. I have been poking around the source; it seems doable but I have no idea about a time frame for it.

[Varti]

Recently I have been working on ZBedic because I wanted to update it and port it to Qt 5 so I can use it on my desktop. Long story short, porting it was too much hassle so I started a clean implementation. That project is called Frasmodic.

Good luck with your new project. If you need any tester for the Linux version, feel free to ask me.

I think it won't be possible to build a Qt 5 version on Zaurus so the best thing to do is to try to update ZBedic. For example, libbedic is statically linked. I can make some improvements to it then build a dynamically linked ZBedic. However, the question is who would use it ? I don't see so many people posting on here saying they are using Sharp ROM and the majority of users seem to use zalarm without GUI (is that correct ?)

Maybe it could be worth a shot to try to run QT5 on Alarmz under a lightweight window manager like Ratpoison. The latest version of the latter has been recently successfully ported to pdaXrom and it is said to be running well. Don't know how fast would QT5 apps run on it though.
I'm at the moment using Alarmz, but once I'll have a stable setup I'm planning to install again Cacko and check its support for Japanese writing, for ZBedic and (if it will be possible to port it) Anki, plus for checking out older apps.
I believe most people here use Alarmz under a shell or pdaXrom; I know of only one user using the stock Sharp's ROM, as he has a 5500.

I also noticed from web history that someone generated a Japanese-English dictionary from EDICT where they mixed kana and kanji keywords together in the index. I have been generating similar dictionary files myself but I separated the kana and kanji versions.  Having a combined index seems very sensible, I think I can adapt my scripts so that is also in my plans.

It would be much more handy to have both kana and kanji displayed for each word result, although for now I only check for kana keywords.

Finally, I am playing with the idea of porting Kanji Nirvana to desktop. I have been poking around the source; it seems doable but I have no idea about a time frame for it.

Never seen it before, I'll definitively try it under Cacko:

http://freecode.com/projects/kani

I have noticed that Freecode is now read-only. I'm worried that it might be taken off-line someday and thus the sources would be lost, we might contact the author and ask him to move the project to e.g. Sourceforge. Also, I have found another project called Nihongo Nirvana which is more recent, is that a fork of Kanji Nirvana or just an updated version of it?

https://sourceforge.net/projects/nihongonirvana/

Varti

[koan]

Thanks for the feedback Varti.

I expect building Qt 5 for Zaurus would take forever, if it is anywhere near possible. It really needs a modern compiler with C++11 support. Even Qt 4 would be a major project. This will take some thinking.

If you are learning Japanese then Kanji Nirvana is an excellent tool for exploring Kanji. It's more of a search engine to find the Kanji you need. Another user adopted the code and started working on updates; he called that project Nihongo Nirvana but his efforts stalled. The writing recognition was provided by KanjiPad* which was updated a few times. My plan is to port to Qt 5 and switch to a newer KanjiPad.

* http://fishsoup.net/software/kanjipad/

[Varti]

I expect building Qt 5 for Zaurus would take forever, if it is anywhere near possible. It really needs a modern compiler with C++11 support. Even Qt 4 would be a major project. This will take some thinking.

I have done a search and found that, luckily, qt5-base and its dependencies are already available as pre-built packages, also for armv5te:

http://de5.mirror.archlinuxarm.org/arm/extra/

Varti

[koan]

I really must get back to working on my dictionary projects and Frasmodic.

In the mean time I updated my release of EDICT for ZBedic. I see there is an EDICT2 which is a richer version of the same dictionary. If it's useful, I will look into adding that extra information to my dictionary files.

[Varti]

EDICT2 seems to be more flexible and useful, plus the older EDICT will be discarded one day, so I believe this new format is the way to go: "This can have multiple kanji forms and readings in an entry, and also has other information such as cross-references, restrictions, etc. and also uses kanji from the extensions in the JIS X 0212 standard."

http://www.edrdg.org/jmdict/edict.html

Varti

[koan]

EDICT2 seems to be more flexible and useful

I agree but the question is how well does EDICT2 fit into bedic format ? - Having lots of annotations for Japanese to English doesn't necessarily apply to other languages and bedic is designed as a multilingual dictionary format.