PRS+: Enhanced firmware for Sony Reader, Folders, Games & more

[readingglasses]

Although I'd love to fantasize about being knowledgeable enough, soon enough, to write it myself, I'll throw out a reminder/suggestion/plea to any enterprising programmer: try to implement a chinese dictionary that can run as an app on this firmware. It would necessarily be vastly different from the dictiolauncher alphabetical list scanner codebase that currently exists (it would need a database with levels).
One could make a chinese dictionary that scanned a list of 10k-40k characters (according to a unicode character code alphanumeric list)...but there could not be a worse way to do it. Even if it would be functionally acceptable when the reader was plugged in and running off mains/usb power, there just could not be a worse way. It would be the easiest to write and run, the most obvious choice for a novice programmer (like myself)...and the worst.

I'll go put in a feature request at prs+.

[kartu]

How does Chinese keyboard look like? They certainly can type on standard US keyboard, turning combination of keys into single glyph, or?

[frank0734]

Originally Posted by kartu

How does Chinese keyboard look like? They certainly can type on standard US keyboard, turning combination of keys into single glyph, or?

One most popular type of Chinese input methods make use of the fact that the pronunciation of each Chinese character can be expressed as some combination of several English letters. Different characters might share same pronunciation, so you usually need to select the one you want from the candidates offered.

This input method is called "Pinyin" and is probably the best choice for a mobile device. However, I do not have an idea if there is an open source input method that can be integrated into PRS+ easily. I'm not an expert in this but I can imagine there should be some standards how this can be done, not only for Chinese but for a lot more languages.

[readingglasses]

Originally Posted by kartu

How does Chinese keyboard look like? They certainly can type on standard US keyboard, turning combination of keys into single glyph, or?

tbh, I had not considered that a chinese dictionary on e-ink must be a type & search dictionary. bc the input method is so rapidly displayed (a list of options that changes as you type more before executing a choice), that the e-ink screen may not refresh fast enough to display clearly or not be a nuissance of black "page turn flashing". My main puzzle is how to create a speedy lookup of a simply selected word on screen. An epub or mobi or other format could do it with a multi level table of contents (root characters as chapters, then a number of different ways of organizing the more complex characters built upon the simple root shapes).

Here's an example of that:
http://hewgill.com/hanzidict/

(I couldn't get calibre to even open that dictionary on my PC but it's structure is hyperlink demonstrated on that site under "dictionary preview")

My idea was it would be a great way to learn unknown characters for which you don't know the pronunciation. I believe there is no input method which exists, without knowing the pronunciation of the character. So typing in a dictionary for a character, not its romanized word, doesn't seem possible, if you don't already know the character.

But here is what I know about chinese keyboards and inputting.


The two photographs on the right side show it perfectly:
http://en.wikipedia.org/wiki/Input_method_editor

http://upload.wikimedia.org/wikipedi.../Ob_gif_ch.gif


http://en.wikipedia.org/wiki/CJK_characters

http://en.wikipedia.org/wiki/Keyboard_layout#Chinese
http://en.wikipedia.org/wiki/Chinese..._for_computers

http://en.wikipedia.org/wiki/List_of...UNIX_platforms



http://en.wikipedia.org/wiki/Smart_Common_Input_Method
http://www.scim-im.org/
http://www.scim-im.org/news/imengine..._tables_0_5_10
http://www.youtube.com/watch?v=7lCN2G5lEzM

http://en.wikipedia.org/wiki/Intelligent_Input_Bus
http://code.google.com/p/ibus/
http://www.youtube.com/watch?v=rvIobt5JpXI
http://www.chinese-forums.com/index....ime-for-linux/

http://en.wikipedia.org/wiki/Google_Pinyin
http://www.youtube.com/watch?v=HVZ2D721Rzo

http://www.m17n.org/m17n-lib-en/index.html

http://en.wikipedia.org/wiki/Uim
http://code.google.com/p/uim/
http://www.youtube.com/watch?v=f-F0p2lqEF0

All of the above is open source. Here's windows doing the same thing, closed source.
http://www.youtube.com/watch?v=3fNeNbmGQrE


which is all very different from russian keyboard on ubuntu
http://www.youtube.com/watch?v=vcVeJWwqh7U

[readingglasses]

Oh yes, and of course, the best open source dictionary database for chinese is

http://cc-cedict.org/wiki/

(which is the basis for the example mobi dictionary in my message above, as the author says on that example hanzi site).

[readingglasses]

So , to answer the question,

using a US keyboard, you might input the character 中 by spelling zhong1 (1 for "high tone"). But as you can imagine, this isn't really necessary for a chinese speaker. If they know the pronunciation of the character, they almost certainly are familiar enough with it to know what it means.
However, it may be useful if they don't know what it means specifically in the context (part of a chain of characters for a longer word) or if they somehow are familiar with it but not its meaning (unlikely).
A chinese dictionary on paper wouldn't be searched by pronunciation anyway, for unknown characters. A chinese speaker would guess what the graphic radical was and try to find the full character under the radical's heading in one of the indexes.
This could be implemented in a computerized dictionary. But they would still have to know how to pronounce at least the simple, radical characters in order to "look up" that chapter/section with all the composites of that radical. Of course, they would know how to pronounce the 200 radical/basic characters. But that still requires some competence.
An ideal dictionary can be of service even to someone who is an incompetent user of the language.

If someone were to speak to them (so not a book/ebook encounter) the word "zhong(high tone)", then they might look up the pinyin pronunciation index alphabetically, or type in what they think they heard in a computer dictionary.

With computerized method, there is also the possibility to draw the character and have the computer do all that graphical recognition and index look up. But it is very ambiguous and error-prone.

Two examples here.
One open source.

http://kiang.org/jordan/software/han...hanzidict.html
(he's not kidding about the stroke order. the program looks up the order of lines, not just the total picture you draw).

(and the pinyin version
http://kiang.org/jordan/software/pinyinime/

One that is not advertised as being open source but is at least freely downloadable (win, mac, linux).

http://www.mandarintools.com/dimsum.html

[readingglasses]

So, as an example for zhong1 中,
you could guess its radical as 'vertical line' and type in "gun3"
or as a box (mouth, 口, type "kou3" 3-"falling and rising tone" ; or upright/proud 囗, type "wei2" 2-rising tone).

Depending on the dictionary maker, one of those (and maybe some others) would be the right answer. The one I just used, was good for the first example (vertical line gun3, plus 3 strokes, "top right corner line" "left vertical line" "bottom horizontal line" to make the box).

Just drawing the character would be neat on a device with a drawing program (like sony prs or now, kobo touch). But it would be error prone and slow as all its potential results would be displayed and thus require a "page refresh/turn" for the user to select the computer's correct guess from its list of guesses.

Here are more examples of the drawing program from above being implemented by other programs:
http://kiang.org/jordan/software/hanzilookup/
look under the "apps" title, scroll down page.


Whereas a straight "select character as text from currently displayed text" could all run in the background and get a guaranteed correct hit, if the character is in the dictionary.

[readingglasses]

But, and this is most essential and the reason why a chinese dictionary would have to have more than one level (either as a database or as a table of contents),
once the hit result was obtained, if you want to be able to "look up" not just a character but a "word" (a modern chinese word is often made up of more than one character/syllable, while only ancient chinese kept most of its words monosyllabic and hence one character only) , the dictionary could then have a second level of "chain compounds" using that character. So zhong+guo is two separate characters and means China. THere is no way to look up "china" and come up with a single character. Zhong means "middle" not china. But when you add "guo/country" it always (or almost always) means "China".

However, a full multi-character word lookup would be a sophisticated, fairly complete dictionary that would be very appealing to more advanced-to-native chinese readers/speakers. Mere novice learners would be satisfied with a character lookup.

Depending on the efficiency of the program, one could have a dictionary lookup to see if it knows a "word" (as a string of chinese "hanzi" characters) but that increases the complexity of the program and the power usage.
And the ccedict dictionary database just lists "words" (multiple characters) as dictionary entries, rather than lists under a character. Which makes for a much longer dictionary to search through. Idk if there's a special attached index that goes with ccedict to make searching easier/faster.

[readingglasses]

http://www.clearchinese.com/chinese-...nary-howto.htm

That's just a full example of the chinese look up characters in a dictionary.
It doesn't include multiple character words...but if you leaf through any chinese dictionary at the bookstore, you'll see a list of multiple character "words" under each character.

[readingglasses]

It's not that nobody wants it or that sony didn't know how. I think they just didn't want to compete with themselves.
They also make electronic dictionaries.

http://www.sony.jp/products/Consumer/DD/index.html

http://translate.google.com/translat...k%2Findex.html

http://www.sony.jp/products/Consumer...8MS/index.html
http://translate.google.ca/translate...n&hl=&ie=UTF-8

During the day included the Sino-Japanese dictionary to help you learn Chinese.
Chunichi DictionaryDictionary Day
Extensive commentary and detailed wording included in the examples and everyday conversation.You can search by typing Pinyin Romanized pronunciation is examined. In addition, it is difficult not to use similar-sounding word, you can search by pressing the button tonal intonation, also helps the conversation.

[kartu]

readingglasses

Let's turn it into some plan for PRS+.
Drawing on the screen to find similar looking chars, while certainly possible, is highly unlikely to be implemented.

1) Keyboard + usual dictionary lookup - will certainly be implemented
2) Keyboard where multiple chars produce single one + dictionary lookup - isn't much more work to do.

Doesn't 2) also cover Chinese? Have in mind, that different lookup methods are actually different dictionaries. (by radical, by number of strokes etc)

PS
It's amazing how inefficient hieroglyphic system is, yet it is still used. Kudos to Koreans for dropping it.

[Pax]

Originally Posted by kartu

readingglasses


PS
It's amazing how inefficient hieroglyphic system is, yet it is still used. Kudos to Koreans for dropping it.

Btw on http://groups.google.com/group/prsp-support/ I opened a thread to ask, if you could implement Korean letters. I provided you with the fonts you asked. How is it going? Does it work? Were the fonts OK?

[kartu]

Hi Pax,

I didn't have time to do it yet, but it will work for sure.

[Arrghus]

A strange idiosyncrasy I've noticed:

When browsing by folder on my PRS-350, authors with names beginning with a lower case letter (I read lots of works published under a pseudonym or username) are pushed to the bottom of the list (i.e. An author with a name like 'acorn' would be listed below a person called 'Zelda', for instance).

Is this intentional?

P.S. Many thanks for a great improvement upon the stock Sony firmware. My particular favourite feature is the clock - it's useful when travelling by train! The book history is also very clever.

P.P.S. I've also posted this message in the PRS+ support group.

[elcreative]

Originally Posted by Arrghus

A strange idiosyncrasy I've noticed:

When browsing by folder on my PRS-350, authors with names beginning with a lower case letter (I read lots of works published under a pseudonym or username) are pushed to the bottom of the list (i.e. An author with a name like 'acorn' would be listed below a person called 'Zelda', for instance).

Is this intentional?

P.S. Many thanks for a great improvement upon the stock Sony firmware. My particular favourite feature is the clock - it's useful when travelling by train! The book history is also very clever.

P.P.S. I've also posted this message in the PRS+ support group.

It's computer sorting... uppercase letters precede lower case in a straight computer sort...