tomoe-devel Mailing List for Tomoe

Hi,

In <6FB...@ya...>
  "[tomoe-devel 245] updated handwriting-ja.xml" on Fri, 14 Nov 2008 14:16:45 -0500,
  Benoit Cerrina <ben...@ya...> wrote:

> I don't know if you guys are interested but to do this I wrote a cocoa  
> tomoe-stroke editor to work on my mac.  I can provide the source code  
> to the project.

Please tell me your SF.net account. I'll add you to the
developers list.

You will be able to put your stroke editor into
  https://tomoe.svn.sourceforge.net/svnroot/tomoe/stroke-editor-cocoa/
or your preferred software name.


Thanks,
--
kou

Hello,
as promized please find an updated handwriting-ja.xml file.  I  
corrected about 160 kanji by adding variants.  I believe that the  
variant is actually the correct way of drawing the kanji but I wanted  
to keep both ways in order to accomodate for
users who do not use the canonical stroke orders (mostly those are  
stroke order changes).

Based on my test using this db with Zinnia and a 0.003 compression  
level (which is not as good as the default for accuracy but better for  
size) and using a reduced db containing only jis 208 level 1 kanji.  I  
went from about 160 kanjis not recognized to 16.
I think this is a good improvement and that all will benefit from one  
of you promoting this.

I don't know if you guys are interested but to do this I wrote a cocoa  
tomoe-stroke editor to work on my mac.  I can provide the source code  
to the project.

Best regards
Benoit

>
The second patch will include the changes already sent but as separate  
variants.  I don't think the last patch I sent should be used as is

> Message: 1
> Date: Wed, 12 Nov 2008 12:31:46 -0500
> From: Benoit Cerrina <ben...@ya...>
> Subject: [tomoe-devel 236]  Face mark + patches for dictionary
> To: tom...@li...
> Message-ID: <1DA...@ya...>
> Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
>
>
> Also I am preparing a second patch to send to the list where I will
> have fixed all the problems I found (I have 110 for now).
> I have decided to make those fixes as new variants of the same
> character as a user could obviously make the same mistake.
>
> Best regards
> Benoit
>
>
>
>

Ok thank you.  As I am mostly interested in zinnia and in editing the  
tomoe dictionary I ended up building my own editor for the mac.
It is cocoa based and not very polished I don't know if you (zinnia  
project) would be interested in the source but if you want I'll send it

Benoit
On Nov 13, 2008, at 12:05 AM, Jens Petersen wrote:

>> file=`echo ja | sed 's,.*/,,'`.gmo \
>>      && rm -f $file &&  -o $file ja.po
>> /bin/sh: -o: not found
>
> I guess you are missing msgfmt from gettext.

I ended up writing my own editor for the dictionary using objectiveC  
on mac.
I can send the source to you guys but it is not very polished as it  
was just for my usage
Benoit
On Nov 13, 2008, at 12:05 AM, Jens Petersen wrote:

>> file=`echo ja | sed 's,.*/,,'`.gmo \
>>      && rm -f $file &&  -o $file ja.po
>> /bin/sh: -o: not found
>
> I guess you are missing msgfmt from gettext.

Point here is that Japanese face mark is generative, i.e. we can
generate any types of face mark with certain combinations of ascii
characters. So, we have a question, who decides which face marks
should be included?   To simplify the policy, I personally think that
the dictionary can contain the characters defined in Unicode and/or
JISX*. Maybe we can have an "extra" dictionary  for Japanese face
mark.

Thanks.

2008/11/12 Mathieu Blondel <mbl...@ru...>:
> 2008/11/12 Taku Kudo <ta...@ch...>:
>> IMHO, we don't want to include face mark into the dictionary, as they
>> are not "character".
>
>
> Actually, there should not be only one dictionary but several.
>
> - All Japanese characters + ascii characters
> - All Japanese characters only
> - Ascii characters only
> - Hiragana only
> - Katakana only
> - Kanjis only
>
> Depending on the application or what the user is looking for, it
> should be possible to switch between them. This improves recognition
> accuracy and computation time. However, this increases the hard disk
> space needed because the same information is repeated in several
> dictionaries.
>
> It would make sense to add "(^ ^)" to "All Japanese characters + ascii
> characters" or "Ascii characters only" but not to the others.
>
> Mathieu
>
>>
>>
>> 2008/11/10 Kouhei Sutou <ko...@co...>:
>>> Hi,
>>>
>>> In <BC7...@ya...>
>>>  "[tomoe-devel 230] strange entry in tomoe dictionary" on Sun, 9 Nov 2008 22:16:23 -0500,
>>>  Benoit Cerrina <ben...@ya...> wrote:
>>>
>>>> I do not understand the entry:
>>>>    <character>
>>>>      <utf8>&#x28;&#x5E;&#x5E;&#x29;</utf8>
>>>>      <strokes>
>>>>        <stroke>
>>>>          <point x="200" y="190"/>
>>>>          <point x="116" y="636"/>
>>>>          <point x="230" y="840"/>
>>>>        </stroke>
>>>>        <stroke>
>>>>          <point x="266" y="350"/>
>>>>          <point x="396" y="246"/>
>>>>          <point x="446" y="360"/>
>>>>        </stroke>
>>>>        <stroke>
>>>>          <point x="553" y="366"/>
>>>>          <point x="666" y="246"/>
>>>>          <point x="763" y="390"/>
>>>>        </stroke>
>>>>        <stroke>
>>>>          <point x="836" y="213"/>
>>>>          <point x="903" y="526"/>
>>>>          <point x="756" y="816"/>
>>>>        </stroke>
>>>>      </strokes>
>>>>    </character>
>>>> in the tomoe dictionary.   What is it supposed to be?
>>>
>>> It's a face mark that is popular in Japan. (^^)
>>>
>>> Why? (1) It shows n UTF-8 characters can be handled as a
>>> logical character in Tomoe. Some input methods can do the
>>> same thing. (2) It's a humor of us.
>>>
>>>
>>> Thanks,
>>> --
>>> kou
>>>
>>> -------------------------------------------------------------------------
>>> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
>>> Build the coolest Linux based applications with Moblin SDK & win great prizes
>>> Grand prize is a trip for two to an Open Source event anywhere in the world
>>> http://moblin-contest.org/redirect.php?banner_id=100&url=/
>>> _______________________________________________
>>> tomoe-devel mailing list
>>> tom...@li...
>>> https://lists.sourceforge.net/lists/listinfo/tomoe-devel
>>>
>>
>> -------------------------------------------------------------------------
>> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
>> Build the coolest Linux based applications with Moblin SDK & win great prizes
>> Grand prize is a trip for two to an Open Source event anywhere in the world
>> http://moblin-contest.org/redirect.php?banner_id=100&url=/
>> _______________________________________________
>> tomoe-devel mailing list
>> tom...@li...
>> https://lists.sourceforge.net/lists/listinfo/tomoe-devel
>>
>
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
> Build the coolest Linux based applications with Moblin SDK & win great prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
> _______________________________________________
> tomoe-devel mailing list
> tom...@li...
> https://lists.sourceforge.net/lists/listinfo/tomoe-devel
>

> file=`echo ja | sed 's,.*/,,'`.gmo \
>       && rm -f $file &&  -o $file ja.po
> /bin/sh: -o: not found

I guess you are missing msgfmt from gettext.

Hello,
thanks for the explanation on face mark characters.  I need to check  
how zinnia is handling it.

Also I am preparing a second patch to send to the list where I will  
have fixed all the problems I found (I have 110 for now).
I have decided to make those fixes as new variants of the same  
character as a user could obviously make the same mistake.

Best regards
Benoit

2008/11/12 Taku Kudo <ta...@ch...>:
> IMHO, we don't want to include face mark into the dictionary, as they
> are not "character".


Actually, there should not be only one dictionary but several.

- All Japanese characters + ascii characters
- All Japanese characters only
- Ascii characters only
- Hiragana only
- Katakana only
- Kanjis only

Depending on the application or what the user is looking for, it
should be possible to switch between them. This improves recognition
accuracy and computation time. However, this increases the hard disk
space needed because the same information is repeated in several
dictionaries.

It would make sense to add "(^ ^)" to "All Japanese characters + ascii
characters" or "Ascii characters only" but not to the others.

Mathieu

>
>
> 2008/11/10 Kouhei Sutou <ko...@co...>:
>> Hi,
>>
>> In <BC7...@ya...>
>>  "[tomoe-devel 230] strange entry in tomoe dictionary" on Sun, 9 Nov 2008 22:16:23 -0500,
>>  Benoit Cerrina <ben...@ya...> wrote:
>>
>>> I do not understand the entry:
>>>    <character>
>>>      <utf8>&#x28;&#x5E;&#x5E;&#x29;</utf8>
>>>      <strokes>
>>>        <stroke>
>>>          <point x="200" y="190"/>
>>>          <point x="116" y="636"/>
>>>          <point x="230" y="840"/>
>>>        </stroke>
>>>        <stroke>
>>>          <point x="266" y="350"/>
>>>          <point x="396" y="246"/>
>>>          <point x="446" y="360"/>
>>>        </stroke>
>>>        <stroke>
>>>          <point x="553" y="366"/>
>>>          <point x="666" y="246"/>
>>>          <point x="763" y="390"/>
>>>        </stroke>
>>>        <stroke>
>>>          <point x="836" y="213"/>
>>>          <point x="903" y="526"/>
>>>          <point x="756" y="816"/>
>>>        </stroke>
>>>      </strokes>
>>>    </character>
>>> in the tomoe dictionary.   What is it supposed to be?
>>
>> It's a face mark that is popular in Japan. (^^)
>>
>> Why? (1) It shows n UTF-8 characters can be handled as a
>> logical character in Tomoe. Some input methods can do the
>> same thing. (2) It's a humor of us.
>>
>>
>> Thanks,
>> --
>> kou
>>
>> -------------------------------------------------------------------------
>> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
>> Build the coolest Linux based applications with Moblin SDK & win great prizes
>> Grand prize is a trip for two to an Open Source event anywhere in the world
>> http://moblin-contest.org/redirect.php?banner_id=100&url=/
>> _______________________________________________
>> tomoe-devel mailing list
>> tom...@li...
>> https://lists.sourceforge.net/lists/listinfo/tomoe-devel
>>
>
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
> Build the coolest Linux based applications with Moblin SDK & win great prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
> _______________________________________________
> tomoe-devel mailing list
> tom...@li...
> https://lists.sourceforge.net/lists/listinfo/tomoe-devel
>

IMHO, we don't want to include face mark into the dictionary, as they
are not "character".


2008/11/10 Kouhei Sutou <ko...@co...>:
> Hi,
>
> In <BC7...@ya...>
>  "[tomoe-devel 230] strange entry in tomoe dictionary" on Sun, 9 Nov 2008 22:16:23 -0500,
>  Benoit Cerrina <ben...@ya...> wrote:
>
>> I do not understand the entry:
>>    <character>
>>      <utf8>&#x28;&#x5E;&#x5E;&#x29;</utf8>
>>      <strokes>
>>        <stroke>
>>          <point x="200" y="190"/>
>>          <point x="116" y="636"/>
>>          <point x="230" y="840"/>
>>        </stroke>
>>        <stroke>
>>          <point x="266" y="350"/>
>>          <point x="396" y="246"/>
>>          <point x="446" y="360"/>
>>        </stroke>
>>        <stroke>
>>          <point x="553" y="366"/>
>>          <point x="666" y="246"/>
>>          <point x="763" y="390"/>
>>        </stroke>
>>        <stroke>
>>          <point x="836" y="213"/>
>>          <point x="903" y="526"/>
>>          <point x="756" y="816"/>
>>        </stroke>
>>      </strokes>
>>    </character>
>> in the tomoe dictionary.   What is it supposed to be?
>
> It's a face mark that is popular in Japan. (^^)
>
> Why? (1) It shows n UTF-8 characters can be handled as a
> logical character in Tomoe. Some input methods can do the
> same thing. (2) It's a humor of us.
>
>
> Thanks,
> --
> kou
>
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
> Build the coolest Linux based applications with Moblin SDK & win great prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
> _______________________________________________
> tomoe-devel mailing list
> tom...@li...
> https://lists.sourceforge.net/lists/listinfo/tomoe-devel
>

Hello,
here is a version of the dictionary with corrections (mostly incorrect  
stroke orders and insuficient number of strokes) to:
355A 及
3037 扱
304F 囲
3052 威
305D 維
3063 違
352E 貴
4C70 矢
303F 或
306D 壱
306F 逸
3121 院
3122 陰
3123 隠
315B 越
3244 可
323F 何
454D 斗
4561 刀
4866 比
332C 階
3354 郭
3356 隔
3B52 子
3358 学
3428 寒
3431 官
3445 甘
3457 閑
3459 陥
345D 丸
3471 奇
3473 寄
3528 季
3532 飢
3554 虐
355B 吸
3566 究
3569 級
356A 糾
357D 享
362B 叫
3665 九
3679 隅
3743 恵
4735 乃
375E 迎
382F 遣
3831 険
3838 幻
3842 限
3849 孤
3925 好
3926 孔
394C 耕
3959 郊
395F 降
397E 込
3A2B 昆
3A2E 混
3A50 歳
3A55 砕
3A5D 際
3A6A 崎

Hi,

In <BC7...@ya...>
  "[tomoe-devel 230] strange entry in tomoe dictionary" on Sun, 9 Nov 2008 22:16:23 -0500,
  Benoit Cerrina <ben...@ya...> wrote:

> I do not understand the entry:
>    <character>
>      <utf8>&#x28;&#x5E;&#x5E;&#x29;</utf8>
>      <strokes>
>        <stroke>
>          <point x="200" y="190"/>
>          <point x="116" y="636"/>
>          <point x="230" y="840"/>
>        </stroke>
>        <stroke>
>          <point x="266" y="350"/>
>          <point x="396" y="246"/>
>          <point x="446" y="360"/>
>        </stroke>
>        <stroke>
>          <point x="553" y="366"/>
>          <point x="666" y="246"/>
>          <point x="763" y="390"/>
>        </stroke>
>        <stroke>
>          <point x="836" y="213"/>
>          <point x="903" y="526"/>
>          <point x="756" y="816"/>
>        </stroke>
>      </strokes>
>    </character>
> in the tomoe dictionary.   What is it supposed to be?

It's a face mark that is popular in Japan. (^^)

Why? (1) It shows n UTF-8 characters can be handled as a
logical character in Tomoe. Some input methods can do the
same thing. (2) It's a humor of us.


Thanks,
--
kou

Hi,

In <674...@ya...>
  "[tomoe-devel 231] correction to dictionary" on Sun, 9 Nov 2008 22:23:39 -0500,
  Benoit Cerrina <ben...@ya...> wrote:

> some of the kanjis in the tomoe dictionary have some stroke order  
> errors.  I am wondering if I should send a correction which contains  
> the correct stroke order as an additional variant (as obviously some  
> people like the person who wrote the dictionary use the incorrect  
> order) or if I should remove the incorrect order and replace it with  
> the correct one?

If you send a patch that corrects them to us, we will apply
it.


Thanks,
--
kou

Hello,
some of the kanjis in the tomoe dictionary have some stroke order  
errors.  I am wondering if I should send a correction which contains  
the correct stroke order as an additional variant (as obviously some  
people like the person who wrote the dictionary use the incorrect  
order) or if I should remove the incorrect order and replace it with  
the correct one?
Best regards
Benoit Cerrina

Hello,
I do not understand the entry:
   <character>
     <utf8>&#x28;&#x5E;&#x5E;&#x29;</utf8>
     <strokes>
       <stroke>
         <point x="200" y="190"/>
         <point x="116" y="636"/>
         <point x="230" y="840"/>
       </stroke>
       <stroke>
         <point x="266" y="350"/>
         <point x="396" y="246"/>
         <point x="446" y="360"/>
       </stroke>
       <stroke>
         <point x="553" y="366"/>
         <point x="666" y="246"/>
         <point x="763" y="390"/>
       </stroke>
       <stroke>
         <point x="836" y="213"/>
         <point x="903" y="526"/>
         <point x="756" y="816"/>
       </stroke>
     </strokes>
   </character>
in the tomoe dictionary.   What is it supposed to be?

Best regards
Benoit

Begin forwarded message:

> From: Benoit Cerrina <ben...@gm...>
> Date: November 7, 2008 6:17:52 PM EST
> To: tom...@li...
> Subject: tomoe data correction
>
> Hello,
> I have made a cross reference of the data from tomoe and the taka  
> database and found lots of kanji entries which probably need  
> corrections (note that I was using zinnia and I really cross  
> referenced the taka db with the version of the tomoe data shipped  
> with zinnia).
> However I haven't been able to launch the stroke editor (it  
> complaints about a missing tomoe import).  I assume this means I  
> must make tomoe itself in order to be able to use the editor.   
> However I have been unlucky instaling tomoe either from the  
> downloadable files
> (configure works but make fails with
> benoit@benoit-desktop:~/tomoe-0.6.0$ make
> make  all-recursive
> make[1]: Entering directory `/home/benoit/tomoe-0.6.0'
> Making all in po
> make[2]: Entering directory `/home/benoit/tomoe-0.6.0/po'
> file=`echo ja | sed 's,.*/,,'`.gmo \
>     && rm -f $file &&  -o $file ja.po
> /bin/sh: -o: not found
> make[2]: *** [ja.gmo] Error 127
> make[2]: Leaving directory `/home/benoit/tomoe-0.6.0/po'
> make[1]: *** [all-recursive] Error 1
> make[1]: Leaving directory `/home/benoit/tomoe-0.6.0'
> make: *** [all] Error 2
> )
> or from the svn source (I can't even get autogen to work).
> Can someone help me get the system running.  If I can I expect to be  
> able to contribute to the quality of the japanese dictionary.
>
> Best regards
> Benoit Cerrina
>
> http://shinkanji.bcerrina.com
> http://padict.sf.net
>
>
> PS:
> if someone else wants to verify the suspicious entries, here is the  
> list I have made of all the jis 208 level 1 kanji in taka which were  
> not correctly recognized by zinnia:
>
>
>
> I have another suspicious set which are the ones for which the  
> recognized kanji is among the 10 best chocies but not the 1st

Hello,
I have made a cross reference of the data from tomoe and the taka  
database and found lots of kanji entries which probably need  
corrections (note that I was using zinnia and I really cross  
referenced the taka db with the version of the tomoe data shipped with  
zinnia).
However I haven't been able to launch the stroke editor (it complaints  
about a missing tomoe import).  I assume this means I must make tomoe  
itself in order to be able to use the editor.  However I have been  
unlucky instaling tomoe either from the downloadable files
(configure works but make fails with
benoit@benoit-desktop:~/tomoe-0.6.0$ make
make  all-recursive
make[1]: Entering directory `/home/benoit/tomoe-0.6.0'
Making all in po
make[2]: Entering directory `/home/benoit/tomoe-0.6.0/po'
file=`echo ja | sed 's,.*/,,'`.gmo \
      && rm -f $file &&  -o $file ja.po
/bin/sh: -o: not found
make[2]: *** [ja.gmo] Error 127
make[2]: Leaving directory `/home/benoit/tomoe-0.6.0/po'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/benoit/tomoe-0.6.0'
make: *** [all] Error 2
)
or from the svn source (I can't even get autogen to work).
Can someone help me get the system running.  If I can I expect to be  
able to contribute to the quality of the japanese dictionary.

Best regards
Benoit Cerrina

http://shinkanji.bcerrina.com
http://padict.sf.net


PS:
if someone else wants to verify the suspicious entries, here is the  
list I have made of all the jis 208 level 1 kanji in taka which were  
not correctly recognized by zinnia: