Improvements of international language support

[myomeriko]

I would like to make a wish list of required features to offer better support of non-English languages for the Kindle 3 and hopefully for the upcoming Kindle 4.
Here is a "dream" list of enhancements.
1. Selectable user interface for various languages: Kindle owners not fluent in English would like to have the menus, fields and messages in their native language. For a start Spanish, German, French and Italian would be welcome. User's Guides in German, French, Italian and Portuguese are already available.
2.Foreign Language Dictionaries: The Kindle comes with two English-English dictionaries (New Oxford American Dictionary and Oxford Dictionary of English). Readers whose native language is not English would welcome translation dictionaries from English to their native language and vice versa. They probably are also looking for translation dictionaries form other languages to English, French, Spanish, Italian etc.
Note that dictionaries are available from third party sources or from the Kindle store.
3. Text to speech of found dictionary term: Many Kindle owners read English books even if his/her native language is not English. Would be useful to add a new text to speech function which would read out the found English word in the correct pronunciation. I have seen this on Chinese e-reader which not only reads out the word but also displays a small image of the "thing". TTS for non-English word pronunciation is probably too complicated.
4. Choice of more fonts: Kindle 3 offers only three typefaces: Serif (regular and condensed) and sans serif. Many users complain about that. This is special true for non-English readers and non-Latin based contentment where users are used to other fonts. Why not let the user install own fonts?
5.Enhanced search options: Search has a number of odd behaviors, even in the English content. First let's see how the find works in Documents (excluding PDF).
a) It does not have wildcards, grep syntax, etc.
b) Multiple words can be entered in the find input area, each separated by a space. The find command will attempt to find all words within a "Location", which roughly equates to a sentence, but not exactly.
c) All symbols (" , -- ' ? ! : etc) are ignored.
d) Common small words are ignored. This differs from language to language. In English "of", "and", "the", "is", "a", "and", "to", "it", "for", "with" etc. In German "die", "der", "das", "daß", "im", "in", "von", "mir", "er", "dich", "sie", "wie", "ist" etc.
e) Searches are case-insensitive. This might be ok for most cases but sometimes an exact search is useful.
f) Characters with accents and diacritical signs can usually be searched by simply ignoring the accent and entering the underlying character (search of German special letters ä=a, ö=o, ü-u, ß=ss and French letter é=e) but this does not work for many other special letters used in other European languages like French, Spanish, Danish etc.
g) Words with apostrophe are not found. This causes problems mainly in languages like French and Spanish where the apostrophe is placed often after the first letter.
6.Better support of non-Latin Languages: Kindle 3 currently official supports Cyrillic, Japanese, Chinese Traditional and Simplified, Korean and Greek characters. I already described some of the problems (Dictionary, Search) but I believe there are other issues I am not aware of. For sure, none of this languages allow typing of notes or search words. I would be happy to receive some comments about this subject.
7. Support of additional Languages: Kindle 3 does not support the following languages: Arabic, Farsi, Hebrew, Urdu (all written from right to left), Thai, Indic scripts (Tamil, Devanagari). There are more but maybe less important. Technically it is possible to develop the support of these languages. Other manufactures have already done it. At least that could supply a "basic" support (without allowing keyboard entry) as they have it with Greek and Cyrillic.
8. Willing to pay for these enhancements? Let's face it. Development of such features cost money and in certain cases we may not be able to get them bundled in the Kindle package. If Amazon develops them, they might want to sell some of them as "extras" as they already do with foreign language dictionaries.

If you like to contribute to any of the subjects, please send me a message before June 30th
Thank you

[odiakkoh]

"Text to speech of found dictionary term"

That would be awesome. It's not enough for me to know how a word is spelled and what it means, I'd like to know how it's pronounced as well.

[tomsem]

If Amazon's next kindle has a touch screen, and virtual keyboard, you can be sure one of the reasons will be to make it more conducive to internationalization and localization. Initially, I would expect localizations for at least German, French, and Spanish, and hope there are no geographic restrictions on which ones you get to choose from.

Text to speech for dictionary pronunciation is problematic for a variety of reasons. Synthesized speech is not accurate enough, and including digital audio of every word defined in the dictionary is expensive in terms of storage requirements. Theoretically the dictionary could include phonetic spellings based on IPA (International Phonetic Alphabet) as indeed the NOAD included with Kindle 3 does (though I don't know if it conforms to IPA). But for other languages, someone would need to add all of those spellings from scratch, which would probably be cost-prohibitive (as would including recorded voices). Seems if you had IPA representation, you could train a synthesized TTS to read it accurately. That would be the best of both worlds but it is not trivial.

I would pay for additional TTS voices in English or other languages that I'm learning.

I would also pay for additional fonts if Amazon sold them.

I would like to pay for dictionaries, but publishers are not making them available. I don't think it is Amazon's fault. Maybe that will change as Amazon moves into more markets internationally. As it is I have collected a few 'public domain' dictionaries that volunteers have produced from PD source material.

BTW I do not see any issues with non-Latin characters in the home screen (pre-K3, yes). Cyrillic for example is okay, I will admit I haven't tried any others. If you are seeing a problem with scripts that are otherwise supported for reading, it might be an encoding issue with the metadata.

Non-latin dictionaries (like Russian) should just work, I keep meaning to send Amazon (another) complaint about this. They are going to have to fix it sooner or later, why not sooner?

[myomeriko]

RE: Text to speech for dictionary pronunciation .
Maybe I did not explain well enough what I was suggesting.
I would limit the text to speech only for English words found in the English dictionaries. I agree that text to speech works not very good in non-English languages. TTS would be helpful for many readers whose native language is not English but frequently read English books. This is probably the major audience using the dictionaries.

[myomeriko]

Re: issues with non-Latin characters in the home screen:
I have a Kindle 3 wich I got two weeks ago from Amazon. I used a Greek book from Project Gutenberg titled "Seven Agains Thebes" with Language metadata set to "Greek". I saved this file on my PC and changed the file name to Greek and transferred it to the Documents folder of my Kindle 3. Instead Greek letters in the title only ??-marks appear. I did the same with a Russian book. Same result than in Greek.
After your comment I made today another test with a txt file which I saved in UTF-8 and named it with 4 letters in Greek and 4 letters in Cyrillic. Surprize: the letters appear correctly in the Home screen. If you want the file I can send it to you. It includes also some more interesting test about Unicode support.

[Alissa]

I believe the tts for multiple languages will be costly, and I don't think it's feasible that Kindle device come with multilanguage tts.

I just want more languages/scripts to be supported as book's primary language. For example, the web browser come with Kindle 3.1 can show Arabic web sites properly, so why not Arabic Kindle ebooks? It should be easy.

[tomsem]

Originally Posted by Alissa

I believe the tts for multiple languages will be costly, and I don't think it's feasible that Kindle device come with multilanguage tts.

I just want more languages/scripts to be supported as book's primary language. For example, the web browser come with Kindle 3.1 can show Arabic web sites properly, so why not Arabic Kindle ebooks? It should be easy.

The browser is based on webkit, and so a lot of the script handling comes 'for free'. But even it has limitations.

Meanwhile, Kindle's reading system was designed to work on puny, memory-constrained systems, and employs a very restricted subset of HTML/CSS. It won't be trivial to modify it for complex-script and right-to-left support (to say nothing about asian-language vertical text layout). But let us hope they are hard at work on it.

That said, I don't think any of the other reading systems support these right now, not even webkit-based iBooks. Except of course what you can do with PDF.

[tomsem]

Originally Posted by Alissa

I believe the tts for multiple languages will be costly, and I don't think it's feasible that Kindle device come with multilanguage tts.

I believe Amazon licenses Nuance Vocalizer technology, which currently supports 40 different languages and dialects. Amazon could license additional voices for sale, I would certainly buy the English dialects if they were reasonably priced (Australian, Scottish, UK, etc.).

[osnova]

These are all great suggestions.

- I also wish they fixed the bug that shows parts of formatting tags when certain two-byte characters (non-Latin) are used.
- The Kindle does not support the pop-up feature for user-created dictionaries in Cyrillic (I know it's weird).
- There is also a search bug where search leads to the wrong destination with certain font sizes (can you believe this?).
- support for RTL languages

[carmelra]

Originally Posted by tomsem

The browser is based on webkit, and so a lot of the script handling comes 'for free'. But even it has limitations.

Meanwhile, Kindle's reading system was designed to work on puny, memory-constrained systems, and employs a very restricted subset of HTML/CSS. It won't be trivial to modify it for complex-script and right-to-left support (to say nothing about asian-language vertical text layout). But let us hope they are hard at work on it.

That said, I don't think any of the other reading systems support these right now, not even webkit-based iBooks. Except of course what you can do with PDF.

Well, you are wrong. iBooks supports RTL languages and there are other reading systems that support it such as Stanza, Overdrive Media Console, PocketBook readers, Calibre, Azardi, EPUBReader and others.