sort of English Duokan

[FethryDuck]

Updated the DXG to 0.5.13.7107 Beta1120 - and think they have fixed some sub-menus, else I have not discovered anything special.

[WS64]

Just curious, is anyone using Duokan already for reading EPUBs?

It's hard to read books in my native German (no Umlauts), and other characters are displayed not at all or really strange which makes it impossible for me to read any books...

So far FBKindle does a way superior job here.

[3rd]

Hello all,

As we know Duokan has problem with national characters from different alphabets. Some of chars are missing, some appear with space after them.
So I wonder if it possible to change something to get national chars.

Changing system fonts of Duokan resolves problem partialy in dictionary: fonts are visible but with additional spaces. However books are still without chars.

In install script of Duokan locales are defined as "zh_CN". Maybe here is the problem. Is it possible to change locales (with adequate files)? Or it's futile?

=============

WS64,
You can prepare one special font with german chars instead of some signs in basic ASCII (up to 127). For example "ß" instead "&" Not the best solution, but better than nothing.

There are languages with 20 and more national chars.

[curstpriest]

The install script contains a zip file that actually installs EVERY LOCALE (it unzips them all onto the kindle, but then deletes them all except the zh_CN one) and uses that locale.

IDK enough about locales but you can mess with it if you do. Also don't know about incompatibilities with the reader. The files are in the word.tgz file or w/e it is called. you can see the file in their update packages. It deletes it after you run the software update to a new version. I guess if you want you could just edit their script to not delete the other locales and then telnet in and set different locales or w/e you want.

http://www.linux.com/archive/feed/53781

locale is just date/time stuff. Actually why the hell is it setting our stuff to Chinese locale =(
Guess it doesn't really matter but.. why?

This doesn't have anything to do with languages..
Pretty sure their reader is broken and only using the standard limited ascii set. Which is probably hard coded into it.

[FethryDuck]

@WS64: "if someone is reading Epub ?" Yes, and some of it is good. I have not checked out FBKindle, so I would not know how they would compare.

@3rd: The character-sets will be solved. I have proposed the "YoCan Reader", as currently I am attempting to make a menu system for several language. Technically it is Duokan, just with a few simple changes, but as it isn't approved of Duokan yet, it was wasn't correct to sail under their flag.

@curstpriest: yup - you seem on top of things here. But for you, I wanted to say that the Uninstall feature has been filed November 11th, but they have not come around to it yet, as far as I can figure out.

You are right about some reader problems, perhaps in relation to character-sets. They are to support Unicode, and this open up a range of issues, as far as I know. ( example text-to-speech, just to say something )

Had no intention of doing this, but after a lot of work on my own English version, perhaps I am a bit biased.

[dtnam]

waiting Duokan for support Vietnamese language

[FethryDuck]

Originally Posted by FethryDuck

@WS64: "if someone is reading Epub ?" Yes, and some of it is good. I have not checked out FBKindle, so I would not know how they would compare.

@3rd: The character-sets will be solved. I have proposed the "YoCan Reader", as currently I am attempting to make a menu system for several language. Technically it is Duokan, just with a few simple changes, but as it isn't approved of Duokan yet, it was wasn't correct to sail under their flag.

@curstpriest: yup - you seem on top of things here. But for you, I wanted to say that the Uninstall feature has been filed November 11th, but they have not come around to it yet, as far as I can figure out.

You are right about some reader problems, perhaps in relation to character-sets. They are to support Unicode, and this open up a range of issues, as far as I know. ( example text-to-speech, just to say something )

Had no intention of doing this, but after a lot of work on my own English version, perhaps I am a bit biased.

EDIT; added link - unicodenstandard

[evilmind]

Originally Posted by FethryDuck

@3rd: The character-sets will be solved. I have proposed the "YoCan Reader", as currently I am attempting to make a menu system for several language. Technically it is Duokan, just with a few simple changes, but as it isn't approved of Duokan yet, it was wasn't correct to sail under their flag.

I dont get it, how exactly can "menus for several languages" help with the fact, that Duocan dont uses unicode characters? Which makes him rather unusable, as its not just matter of menu looks, we cant read books too - which is kinda crucial for ebook reader... ;-)

[3rd]

curstpriest,

In the file you've menshioned (word.tar) is folder named "zh_CN.utf8". There are LC_* files. The script use command "localdef" to define locales first and only then copies that folder into /usr/lib/locale/. So spacial LC_* files are prepeared only for zh_CN. I know too little about meaning of LC_* files, but maybe LC_CTYPE and LC_COLLATE files have effect on charsets.

I was unable to make usbnet connection with my Kindle running Duokan. But I think that I can set locales in install/update scripts without connecting to Kindle directly, what bothers me is that folder with LC_* files.

Eventually problem could lie not in locales but somewhere in Duokan software alone.

=========

FethryDuck,

Translating menus is good idea, but it is hard to make translations with invisible chars ;-)

[MrZeroo00]

It has already been said on this topic, but French characters (like é, è, ç, etc..) are rendered in a font different from the one used in the text, it makes Duokan unusable for all French texts, any solution ?

[curstpriest]

Originally Posted by MrZeroo00

It has already been said on this topic, but French characters (like é, è, ç, etc..) are rendered in a font different from the one used in the text, it makes Duokan unusable for all French texts, any solution ?

http://en.wikipedia.org/wiki/ISO/IEC_8859-1

It's dumb, and assumes all documents are encoded with a certain set or something, I'm not sure what set it uses as default to display text when it can't read the encoding from the file (or what it assumes everything is as default). 8859-1 is pretty complete and by far the most used worldwide, but perhaps it's not using this.
Anyways, you'll see 2x-7x characters properly on the chart, regardless of encoding, but AX-FX will look strange (enlarged, wrong font) because the document is probably encoded in ISO 8859-3 or something.

It could be using unicode, but I kind of doubt that, but I don't know enough about encoding to tell you.

Characters NOT on the chart will just be omitted, or worse, characters next to the omitted ones will also be missing because the reader will error and not be able to render the word.

The simple way to fix it, is to tell duokan developers to support all 15 encodings (read it from the file when it's opened, and render text with that character set)

I could be wrong, are your characters not in this list (image above) what exactly is happening?

[3rd]

I've tried to change locales in update script but without visible effect. So it looks like the problem with Duokan software itself.

_______________

MrZeroo00,

You can use the same font for Chinese and English, also your custom fonts. And then all fonts will be the same. However non-english chars are with additional space.

I may be wrong, but there's no such font as "ç".

_______________

curstpriest,

the best solution is to use Unicode, and if you want custom encoding you need only custom font with proper glyphs.


For me looks like Duokan uses UTF-8, but then cuts most of national fonts somewhere in Duokan code. Remaining national fonts are taken from Chinese font-file with extra space.
There is some difference between books and dictionary. Dictionary sees most of the fotns from 256, however books not. Fonts from 160 to 255 are without extra space in dictionary, but only those permitted by Duokan code.

[FethryDuck]

hi all,

Invisible characters ? NOPE, it is a ZEN thing - just practice your meditation and wait for the big AHA !

Perhaps the version was a bit premature, as I have found several errors. There are some mess with character sets, granted. And what type of books, epub, mobi, pdf, etc. Mobi is not supported by the reader yet, and epub has a few issues. It may be beta code for Chinese, while the English version was just released. So it is bound to be some issues.
French, which others ? Any sample texts ?

BTW. OCR - does on occasion produce some real weird results. For example Gutenberg texts. I have seen example of it - so...some references and samples would be nice. And printscreen is shift+space.

[curstpriest]

Another thing - when you read an EPUB, it converts it to HTML (or XML), I suggest you open that up (it's in the DK_MDB folder or w/e) and look at it, see if it states the encoding (it might even be UTF-8 - that would be bad i think) and see if the text looks alright in the converted html document. If it looks bad in the converted document, well then that's part of the problem

Also locales are only responsible for date/time/currency format, it would not have any effect on text rendering. i don't even know why they make an entry for it.

[FethryDuck]

Originally Posted by curstpriest

Another thing - when you read an EPUB, it converts it to HTML (or XML), I suggest you open that up (it's in the DK_MDB folder or w/e) and look at it, see if it states the encoding (it might even be UTF-8 - that would be bad i think) and see if the text looks alright in the converted html document. If it looks bad in the converted document, well then that's part of the problem

Also locales are only responsible for date/time/currency format, it would not have any effect on text rendering. i don't even know why they make an entry for it.

Oh, you are really digging in - that was good thinking ! They do have some other date/time formats, so they need it. ( I am guessing some of course, as I am sure much would be much more obvious if you read their languages )

Thank you