Hello,

thank you for replying.  Perhaps I should have just asked if there was a
Japanse IME for Squeak.

Your page said that you were using the operating system IME, but I couldn't
get the Windows 2000 Japanese IME to input into Squeak.

I'm doing some searches on SKK.  So far I haven't seen how to install SKK in
Squeak.  I've see dictionary files, and links to images for pocket machines
(I think) and to Unix input servers...  My Japanese is very poor, so no
doubt I'm misunderstanding the explanations.

Do you know where I can find a general purpose squeak image with an IME
installed or where I could find smalltalk code for the IME or, better yet,
is there there a project in a squeak-map server?

While I'm sure it would help my understanding of Japanese to write an IME
and it would be useful to have a bilingual interface:

1. I really only need a Japanese IME.
2. I doubt I could write an IME that works as well as one written by someone
who knows more Japanese than I do.  I'm in my first Japanese class in fact.
3. I'd have to figure out where to hack into the Squeak source (I've been
working on that)
4. I'd have to figure out how to deinflect the Edict dictionary and add
correct inflections for all of the verbs and adjectives into the input
dictionary (and perhaps there's more I don't know about).  That's possible
since all of the verbs are tagged with extra information.
5. I'm sure that better Japanese speakers can solve problems like segmenting
Japanese text better than I can (finding word boundries isn't easy).  My own
interface would have rely on people hitting the space bar or return between
almost every word or it would have to find every possible segmentation and
give the user a choice between them all.  It would take time to write a user
interface that could do all this.
6. It's better not to write software that's incompatable with what's already
being used.  If Squeak users already have an IME, I should use that, and add
features to that if necessary

I've written an audio flash-card program in Squeak for Japanese students and
bilingual input for visual flash-cards would be a nice feature.  I can see
that SKK dictionary page does incude one of the Edict Japanese/English
dictionaries, by the way.  Maybe there's already some sort of bilingual
option in SKK.  If not, perhaps I can add one in the future.

Except for cutting text out word processor that has an optional Shift-JIS
clipboard feature and pasting it into Squeak, I couldn't find any way to
enter Japanese characters.  For some reason even cutting and pasting
Japanese text out of Microsoft Word and most other programs didn't work -
apparently it wasn't Shift-JIS in the clipboard or wasn't tagged properly.

I would appreciate any help you could give me about how to install SKK.
Maybe there's already an IME installed in the SqueakNihongo image and I just
didn't figure out how to invoke it.

Thank you for your help.

Joshua Scholar

> ----- Original Message ----- 
> From: Yoshiki Ohshima
> To: Joshua Scholar ; The general-purpose Squeak developers list
> Sent: Friday, November 26, 2004 10:59 PM
> Subject: Re: Where would I post a BIG project?
>
>
>   Hello,
>
>> I'm thinking of using the Edict dictionaries (
>> http://www.csse.monash.edu.au/~jwb/j_edict.html ) to make a Japanese IME
>so
>> that you can enter Kanji (chinese derived) characters from the keyboard.
>>
>> The Edict dictionaries are Japanese/English dictionaries, so other
>features
>> are possible, like Japanese entry for english speakers that shows English
>> meanings as well as choices of Kanji characters.
>
>  That would be helpful.  Some people did a Squeak-only
> kanji-converter based on SKK, but showing the English meaning is not
> something an ordinary Japanese user would imagine to implement ^^;
> (i.e. would want to have)
>
>> What I'm wondering is this.  I can't post a change set that's 20
megabytes
>> (or however large the dictionaries would be after loading), and I can't
> tell
>> the users to just download the dictionary from the Edict site because
it's
>> in EUC-JP format and needs to be converted to UTF-8 before it can be
> used...
>
>   Not really.  Do you notice EUCJPTextConverter class?  If you write
> a code snipplet something look like:
>
>   file := FileStream readOnlyFileNamed: 'file'.
>   file converter: EUCJPTextConverter new.
>   contents := file contentsOfEntireFile.
>
> it should give you proper "contents".
>
> (However, I agree that it would be easier for the users to have only one
> file to download.)
>
>> I have a word processor that can do that conversion, but I can't expect
>> users to have that.
>
>   Squeak only solution would be nice.  Something like FreeWnn
> connectivity would be nice, too.
>
>   Thank you!
>
> -- Yoshiki
>