Quote:
Originally Posted by reprep
still waiting for the update which
1) shows latin characters as they should be (ç,ş etc)
2) doesn't put space after apostroph. Seriously that's one major flaw keeping me still at the original amazon firmware.
|
Me too ! - still waiting
Anyway, it cannot be that bad ? Not sure what language you use, but the "ç" works in Portugese, or think it worked. Ok, downloaded a book and was surprised by the result, as it did not show that character.
Strange... looking again for that previous book. I don't speak Portugese, and did only select some random sentence. Google translate seem to say:
"I owe everything, life, education, the position we now have in society.
And you do not ignore the sacrifices that such education has cost you. "
In Portugese:
"Tudo vos devo, a vida, a educação, a posição que ora tenho na sociedade.
E vós não ignoraes os sacrificios que essa educação vos tem custado."
And here is what Duokan shows:
"Tudo vos devo, a vida, a educao, a posio que ora tenho na sociedade.
E vš®s no ignoraes os sacrificios que essa educao vos tem custado."
That is pretty clever, huh ? It simply ignore such letters as "çã", dumping U+00E7 and U+00E3. But then it figure to insert some extra code ( where from is a mystery ), as in the next sentence it replaces U+00F3 with U+0161,U+00AE ! Does it believe it is a registered trademark ?! - is it an example of Duokan artificial intelligence ?
How could I ever have believed it had displayed Portugese ? Well, look at the following printscreen copies !
( I did cut out the above sentences, instead of showing the whole page )
The first - Screenshot-13.png - show the above missing characters. But wait, what did happen to the AI-insertion ? -now it seem correct ?
The second - Screenshot-12.png - is the reason why I did believe it once did display things correct. Checked it byte by byte, and this is actually a printscreen of the same book ! ( I changed fonts, just to check if that had anything to say )
The strange results I got, is by having the book in /documents folder ( which display it correctly ) and by having it in the /DK_Documents folder ( which show missing characters )
Screenshot-11 shows the text marked to digest. The output of the file under /DK_Daochu folder ( they could name it DK_Digest, as well as corrected the menu error "dict" instead of "digest" for exporting it ).
digest-file:
--------------------------------------------------------------------
pg22044
[Digest1]
Tudo vos devo, a vida, a educao, a posio que ora tenho na sociedade.
E vš®s no ignoraes os sacrificios que essa educao vos tem custado.
---------------------------------------------------------------------
As seen, it is dropping the characters which makes it impossible to have it looked up in the dictionary. Furthermore, the bytes offset the character positions for marking the word. The Portuguese input does not hold these characters, so the settings of input language is bogus.
It is too stupid - it could all have been corrected a year ago with some source available, and some joint effort.
