Conversion from Simplified Chinese to Traditional Chinese?

[Nobumi Iyanaga]

Hello,

I am wondering if there is any "authorised" conversion table from the
so-called "Simplified Chinese" characters to "Traditional Chinese"
characters.

I know that the correspondences between the two character sets are
not simple. There are cases in which there is one-to-one
correspondence (for example 东 and 東), but there are many other
cases in which this is not true (for example 台 can be either 台,
檯, 臺, and 颱, etc.).

So, the ideal would be a conversion table which contains all the
possible conversions of given characters.

Thank you in advance for any insight.

Best regard,

Nobumi Iyanaga
Tokyo,
Japan

[Magnus Lewan]

I doubt that there is any officially "authorised" conversion. It seems just like the kind of thing that people may not make up their minds about.

The free tables I can come to think of right now are the CEDICT dictionary, where I think both variants are taken up. I do not know if they are stored in any usable way though, and it is not very complete.

Then there is OpenOffice, that has a tool to convert between the two. As OpenOffice is open source, I imagine one could get hold of the table used for conversion if one downloads the source code. I do not know where in the source it would be though.

Good luck!
Magnus

[Kerim Friedman]

Google translate does Simplified <> Traditional conversion and I think
it has an API? Although an API isn't the same thing as tables, it
might do the trick?

kerim

> --
> You received this message because you are subscribed to the Chinese Mac group.
> For answers to frequently-asked questions, visit http://www.yale.edu/chinesemac
> To start a new topic, send a new message to chine...@googlegroups.com
> To unsubscribe, send a message to chinesemac-...@googlegroups.com
> For more options, visit http://groups.google.com/group/chinesemac

[Edward Lipsett /t]

on 09/12/14 11:17, Kerim Friedman wrote:

> Google translate does Simplified <> Traditional conversion and I think
> it has an API? Although an API isn't the same thing as tables, it
> might do the trick?

I can't imagine how it would be possible to implement SC-to-TC without
contextual parsing. The other way is easy because many reduce to few, but
the only way a computer can guess which of many possible TC characters is
correct would be to understand the sentence.
...which is pretty unlikely...

Computers might be able to make a good guess, and could certainly pop up a
list of candidates, though!


--
Writing has laws of perspective, of light and shade, just as painting does
or music. If you were born knowing them, fine. It not, learn them. Then
rearrange the rules to suit yourself.
- Truman Capote
--
Edward Lipsett
translation€@intercomltd.com
www.kurodahan.com

[Kerim Friedman]

Unlike traditional translation engines, Google Translate does not work
by "understanding the sentence." Rather, Google takes a brute force
approach which leverages its large statistical database. It
essentially makes an educated guess on the likelihood of a certain
word being a translation for another word based on context, rather
than knowledge of syntax.

Kerim

[Edward Lipsett /t]

on 09/12/14 11:36, Kerim Friedman wrote:

> Unlike traditional translation engines, Google Translate does not work
> by "understanding the sentence." Rather, Google takes a brute force
> approach which leverages its large statistical database. It
> essentially makes an educated guess on the likelihood of a certain
> word being a translation for another word based on context, rather
> than knowledge of syntax.

Yeah, I know... But a statistical database does not equal context.
It merely means playing the percentages.

Which is what I said: Computers can make a good guess.

--
A novel is balanced between a few true impressions and the multitude of
false ones that make up most of what we call life.
- Saul Bellow, Nobel Prize acceptance speech

[Jean Soulat (Smarthanzi.net)]

Hello,

A good reference is the Unihan file from Unicode.org. This is a text
file used by many applications or websites. It probably relies on
official Chinese standards but I do not know them.

I show Unihan correspondences in www.smarthanzi.net, use the full
popup for a quick view.

Best regards,
Jean

[Weiyun Yu]

There's a simple TC-SC conversion utility called isDrop written by Jeff Ye
some years ago that does this job in a simple manner. It hasn't been further
developed for a while and is at v1.1(12) but still works under 10.6.2.
Although not a conversion table, but it does a pretty decent job for basic
needs.

--


Kerim Friedman at oxu...@gmail.com wrote on 14/12/09 13:17 :

[TenThousandThings]

I don't know what you are planning, but you should also be aware that
there is an effort underway to both add to the list of official
simplifications for general-use [通用] characters and also to undo or
change a few (maybe five or six) of the most problematic of the
overlapping simplifications. In other words, unless I'm mistaken, a
few of the old TC/SC and SC/TC conversions are likely to change in the
near future. I believe the current proposed list contains 8300
characters. See here:

http://www.china-language.gov.cn/gfhzb/

http://www.china-language.gov.cn/doc/zb2009.pdf

The last time they did this was 1988, which produced a list of 7000
characters. [现代汉语通用字表]:

http://www.china-language.gov.cn/gfbz/

http://www.china-language.gov.cn/gfbz/shanghi/014.htm

[TenThousandThings]

Here is another location that has mostly the same documents, but not
in electronic-text form (i.e., they are just scans of the printed
documents):

http://www.china-language.gov.cn/wenziguifan/

So these should be used together with the electronic ones:

http://www.china-language.gov.cn/gfbz/

Compare, for example,

http://www.china-language.gov.cn/wenziguifan/managed/002.htm

with

http://www.china-language.gov.cn/gfbz/shanghi/002.htm

...

[Nobumi Iyanaga]

Hello,

Thank you very much all who replied to my query.

The main reason for which I would like to have a table of
correspondences between Simplified Chinese characters and Traditional
Chinese characters is that..., I have to edit an international
scientific journal in the field of East Asian traditional cultures
and religions (mainly China, Korea and Japan, but other cultural
areas too).

In some Japanese journals in this kind of fields, the articles are
edited so that simplified Sino-Japanese characters (say such as 国)
are systematically "converted" into their traditional forms (such as
"國"). I think this is a good practice, because it allows a certain
standardization, avoiding confusion. In Japanese, this is rather
simple, because in almost all cases, the correspondence between
simplified forms and traditional forms is "one-to-one" correspondence
(and there are good tables of correspondence, from which it is rather
easy to write scripts of conversion). But I think applying the same
kind of conversion between Simplified Chinese and Traditional Chinese
is not simple at all, because in many cases, the correspondences
depend on contexts and meanings.

I could make a table editing somehow the "Unihan_Variants.txt" that I
could find in Unihan database files. It gives a list like this:

㑩 儸
㓥 劏
㔉 劚
㖊 噚
....
临 臨
为 為, 爲
丽 麗
举 舉
么 幺, 麼, 麽
义 義
....

etc.

I think that with this table, it is possible to perform basic
conversions, and propose candidates for other possible conversions.
It would be perhaps interesting to publish this table in a web page
(I can do this if there are any interest into it).

On the other hand, although this is not directly a table of
correspondence between Simplified Chinese and Traditional Chinese,
there is an interesting table of correspondence between Simplified
Chinese (GB 2321) and Japanese characters (JIS X 0208), which you can
find at:
<http://www2u.biglobe.ne.jp/~kitabo/softwares/KCTTBLs.lzh>

It contains a table like this:

1602阿
1603埃
1604挨
1605
1606
1607哀
1608皚
1609癌
1610藹
1611矮
1612艾
1613碍礙
......

I think using this table, it would be possible to write a more
comprehensive table of conversion. The problem is that the first four
digits represent the "KuTen code" for each GB 2321 character, and I
don't know how to convert this code to GB characters...

There may be other ideas.

I would appreciate any further insights.

Thank you in advance.

[Eric Rasmussen]

On Tue, Dec 15, 2009 at 9:54 PM, Nobumi Iyanaga <n-iy...@nifty.com> wrote:
> I could make a table editing somehow the "Unihan_Variants.txt" that I
> could find in Unihan database files. It gives a list like this:
>
> 㑩       儸
> 㓥       劏
> 㔉       劚
> 㖊       噚
> ....
> 临       臨
> 为       為, 爲
> 丽       麗
> 举       舉
> 么       幺, 麼, 麽
> 义       義
> ....
>
> etc.
>
> I think that with this table, it is possible to perform basic
> conversions, and propose candidates for other possible conversions.

As you know, the universe of traditional variants is an entirely
different problem from converting from SC to TC. You have chosen two
very good examples that illustrate this:

> 为 為, 爲

为 = [為 or 爲] is a question of variants. This is not a SC-to-TC
conversion problem. The meaning of 为 is not ambiguous in simplified
Chinese. 為 and 爲 are two variants of the same character. Whether to
use 為 or 爲 in traditional Chinese is a question of publishing norms,
not language. After that choice is made, it is a simple one-to-one
conversion.

> 么 幺, 麼, 麽

This is one of the most confusing problems in conversion, largely
because 幺, 么, and 麽 all currently exist in simplified Chinese. Note
that 麽 is likely to be eliminated (i.e., officially simplified to 么)
in the upcoming revision of simplified Chinese I mentioned earlier
<http://www.china-language.gov.cn/doc/zb2009.pdf>.

In the current state of simplified Chinese (i.e., the 1988 list of
7000 characters
<http://www.china-language.gov.cn/wenziguifan/shanghi/014.htm>),
however, 麽 is the simplified form of 麼 only when it is pronounced mó,
which pretty much only occurs in one word, yāomó. Compounding the
problem is the fact that the first character in the word is yāo, which
can be written either 么 or 幺 in traditional Chinese, but it is
officially simplified as 幺 in simplified Chinese. So in simplified
Chinese yāomó is always 幺麽.

Here are two sets of tables I did in 2006:

http://homepage.mac.com/chinesemac/SC-TC.txt
http://homepage.mac.com/chinesemac/TC-SC.txt

I don't remember why or how I made them. I probably asked Wenlin to
convert all of GB2312 to TC and all of Big Five to SC, and then made
lists of the instances where it asked me to resolve ambiguities.

Eric

[Nobumi Iyanaga]

Hello Eric,

Thank you very much for your examples and the "ambiguous
correspondences" lists. As I am not fluent in Chinese (at all...; I
can only read Buddhist Chinese texts), it is very difficult for me to
understand/distinguish different meanings of characters, especially
when they are in compounds and/or when they are in Simplified Chinese.

Your lists will help me to distinguish cases in which I will able to
do simple conversion, and others for which I will have to ask the
help of people specialized in Chinese.

How much does Wenlin cost? Perhaps I should buy it...?

Thank you again.

[Joe Wicentowski]

Hi Nobumi,

> How much does Wenlin cost? Perhaps I should buy it...?

Wenlin is USD200; see their order page at http://wenlin.com/order.htm for international orders.  Unfortunately the demo doesn't contain the conversion functionality.  See below for screenshots showing the steps to convert Eric's example character, 么 > 麼  (I hope the images come through on the list).

Joe

Fig. 1: A document containing a single character

Fig. 2: Selecting Edit > Make Transformed Copy... brings up this dialog.  

Fig. 3: Choices available in the Make Transformed Copy dropdown menu.

Fig 4: Dialog describing transformation

Fig. 5: Transformed text, showing alternate characters. Click on either character to select it.

Fig. 6: Final transformed text.

[Eric Rasmussen]

On Wed, Dec 16, 2009 at 8:33 PM, Joe Wicentowski <joe...@gmail.com> wrote:

example character, 么 > 麼  (I hope the images come through on the list).

Fig. 1: A document containing a single character

Fig. 5: Transformed text, showing alternate characters. Click on either character to select it.

Someone might ask, why isn't 麽 an option here? The answer is because in this context (i.e., the context of converting 么 from SC to TC) 麽 and 麼 are variants. So Wenlin has chosen one of them, just like the choice it makes between 為 and 爲 when converting 为.

Two other clarifications:

[1] To complete my discussion of yāomó, note that in the 2009 draft of the revised simplification list, it is to be written 幺么 instead of 幺麽. Since 麽 in simplified Chinese is itself a simplification (of 麼), the new way of writing it will be a simplification of a simplification!

[2] On traditional-Chinese variants, we who use Mac OS X are very lucky to have an excellent tool, the Unihan Variant Dictionary:

http://www.ideographer.com/unihan/

This goes well beyond the Unicode list. It is by no means complete (I have a file on my desktop that is a running list of variants I have encountered that are not in it), but the author has been disciplined about what to include, relying on authoritative secondary sources, so there are few (if any) mistakes in it.

Eric

[Joe Wicentowski]

> [2] On traditional-Chinese variants, we who use Mac OS X are very lucky to have an excellent tool, the Unihan Variant Dictionary:
> http://www.ideographer.com/unihan/
> This goes well beyond the Unicode list. It is by no means complete (I have a file on my desktop that is a running list of variants I have encountered that are not in it), but the author has been disciplined about what to include, relying on authoritative secondary sources, so there are few (if any) mistakes in it.

Indeed! I should've mentioned it earlier. I am happy to have
originally proposed the Unihan Variant Dictionary to Ajie 阿傑, aka
Ideographer, over a cup of coffee in Taipei a few years ago. He did a
great job. He kindly accepted the introductory text I wrote for the
program's homepage. His background in linguistics made him the
perfect developer for this.

Ajie also writes an excellent blog on language issues and/or the Mac:
http://www.ideographer.com/articles/.

Joe

[Jean Soulat (Smarthanzi.net)]

I have added a conversion tool in www.smarthanzi.net
- Copy a Chinese test
- Click on "s/t test" (instead of Confirm)
- The website provides the simplified / traditional correspondences
from the chosen dictionary: CEDICT, HanDeDict or DDB.

At the moment it is only a test, please let me know if it is helpful.

Jean

[Nobumi Iyanaga]

Hello Jean,

I tested it. For me, it is tremendously useful. Unihan Variant
Dictionary is very good, but it only gives variants of single
characters, while your tool allows to convert many characters at
once, and gives not only single characters meanings, but compounds
meanings too. Very useful for non fluent people in Chinese like me.

Thank you VERY much!

[Jean Soulat (Smarthanzi.net)]

Thanks Nobumi,

I shall integrate it correctly in the website.

Best regards,
Jean

[Nobumi Iyanaga]

Hello Jean,

After I sent my previous message, I thought it would be even better,
if you could implement two kinds of output, one simplified, and the
other, the complete, which is in fact the current output. The
"simplified output" would display the result of conversion in the
same way as the input. I mean, if the input is a one line string, the
result would be displayed in one line. In case there are more than
one candidate for a character, the two or more characters would be
displayed for example as "a［b/c］". For example, if the input is:
定襄县中霍村东周墓发掘报告
the "simplified output" would be:
定襄縣中霍村東周［週］發［発／髪／髮］掘報告

At the same time, the current output, with its analysis of the words,
pronunciations, meanings of words and of each character, etc. -- all
of this is VERY useful for someone as me, who is not good in Chinese
-- would be displayed as well.

----

By the way, the above example gives currently no output for "中霍
村". On the other hand, it gives as possible candidates of "周",
"週" and "賙". But the latter character is not given as a candidate
for "周" according to Unihan Variant Dictionary. Is there any reason
for this "賙"?

Anyway, as it is now, your website is extremely instructive. Thank
you VERY MUCH for this wonderful tool.

[Jean Soulat (Smarthanzi.net)]

Hello Nobumi,

1) Your proposal is a good idea. I need to make some tests to check
what exactly is possible.

2) No output for "中霍 村": yes, because I do not show words where
traditional and simplified are the same. At the moment it is not
clear, it should be ok with the two text outputs.

3) “周”, “週” and “賙”
I use several different sources for characters (Unihan) and for words
(CEDICT, HanDeDict and DDB).
For each word, CEDICT and HanDeDict provide both variants (traditional
and simplified).

Here is the CEDICT entry:
賙 周 [zhou1] /bestow alms/
So CEDICT provides周 for simplified while Unihan says 赒 (see for
instance Confirm + full popup). I use the CEDICT information “as is”.
It would be technically difficult to compare and I do not have the
Chinese background to make a decision. The mistake – if any – should
be corrected on the CEDICT MDBG website.

Best regards,

Jean Soulat
Rennes
France

[Nobumi Iyanaga]

Hello,

I wrote a very simple Perl script which can convert from simplified
Chinese to Traditional Chinese. For all cases there are more than one
candidates, it generates a string like "candidate_1/candidate_2/...".

This is certainly not perfect, but I think it can be of some use for
some researchers (or other people)...

I uploaded it at:
<http://www.bekkoame.ne.jp/~n-iyanag/researchTools/
simplified2traditional.html>

I just wrote it, and not tested it extensively. So please use it with
caution. Any feedback will be appreciated.

[Jean Soulat (Smarthanzi.net)]

Hello Nobumi,

I have integrated the conversion facility in www.smarthanzi.net
It includes your proposal for the text output: depending upon the
case, the text is converted from simplified Chinese to traditional
Chinese or from traditional to simplified.

Best regards and happy new year,
Jean

[Nobumi Iyanaga]

Hello Jean,

Thank you for your message, and the new implementation. Yes, it will
be so useful for me. I appreciate so much your efforts!

Thank you again.

[Michael Engel]

Has anybody still these files ?
It seems as if the page is not available anymore ?
Thanks
Michael

2009/12/15 TenThousandThings <hello....@gmail.com>:

[Jean Soulat (Smarthanzi.net)]

Michael,

The page is still available:
- Go to www.smarthanzi.net
- Copy your text and confirm.
- Then the choice "Standard" (= default view) or "Simpl/Trad" (=
conversion simplified-traditional) is available in the toolbar - or
traditional to simplified if you entered a traditional text.

It will be improved for long texts in a few weeks, the result is
currently not very easy to copy into a word processor when it is too
long.

Regards,
Jean

[peppermint patti]

Here's another site that converts between simplified and traditional Chinese characters.  However, I have only use it to convert paragraphs at the most.  I have never really pushed it to see what the maximum length of string it can handle.

簡繁轉換 http://www.j4.com.tw/big-gb/

--

[wwg...@gmail.com]

There are some APIs out there you can use to dynamically translate your content, such as this one I have been using: https://ts-api.com

On Sunday, December 13, 2009 at 12:20:46 AM UTC-8, n-iyanag wrote:

Hello,

I am wondering if there is any "authorised" conversion table from the  
so-called "Simplified Chinese" characters to "Traditional Chinese"  
characters.

I know that the correspondences between the two character sets are  
not simple. There are cases in which there is one-to-one  
correspondence (for example 东 and 東), but there are many other  
cases in which this is not true (for example 台 can be either 台,  
檯, 臺, and 颱, etc.).

So, the ideal would be a conversion table which contains all the  
possible conversions of given characters.

Thank you in advance for any insight.

[Jens Østergaard Petersen]

Hi Nobumi,

What you are looking for would be tables like the one-to-many code mapping tables maintained by the CJK Dictionary Institute <http://www.cjk.org/cjk/c2c/c2centry.htm>. I have always thought that their article at <http://www.cjk.org/cjk/c2c/c2cbasis.htm> served as the best introduction to the problems involved. Their tables are certainly proprietary, but one can extract one-to-many correspondences from the cdphanzi files at 漢字構形資料庫 <https://cdp.sinica.edu.tw/cdphanzi/>. I worked with an earlier version years ago, importing the tables into a FileMaker database – but now I have trouble figuring out what exactly I did then …. However, I attach a list I have managed to export. It contains more than 正體字 equivalences, but if you are interested, I could work some more on this ….

Cheers,

Jens

--

--
You received this message because you are subscribed to the Chinese Mac group.

For answers to frequently-asked questions, visit http://www.chinesemac.org

To start a new topic, send a new message to chine...@googlegroups.com
To unsubscribe, send a message to chinesemac-...@googlegroups.com
For more options, visit http://groups.google.com/group/chinesemac

---
You received this message because you are subscribed to the Google Groups "Chinese Mac" group.
To unsubscribe from this group and stop receiving emails from it, send an email to chinesemac+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/chinesemac/f704fd27-dc6c-4539-bd3c-31f260ea0180%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Ho Ian]

Ministry of Education. Taiwan, Republic of China.

標準字與簡化字對照手冊

https://depart.moe.edu.tw/ed2500/News_Content.aspx?n=1DD9A37B3B3857B4&sms=174DD55894170669&s=1DE13940BE5275D9

n-iyanag於 2009年12月13日星期日 UTC+8下午4時20分46秒寫道：

[Jens Østergaard Petersen]

This is authoritative, true, and fine as a beginning, but not comprehensive enough, since it fails to note many cases where what is registered as a simplified character can also be an unsimplified character. A few examples from the first pages, with the BigFive code point added):

 

丰-豐 (丰 A4A5)

厂-廠 (厂 C944)

万-萬 (万 C945)

宁-寧 (宁 C972)

Jens

--
--
You received this message because you are subscribed to the Chinese Mac group.
For answers to frequently-asked questions, visit http://www.chinesemac.org
To start a new topic, send a new message to chine...@googlegroups.com
To unsubscribe, send a message to chinesemac-...@googlegroups.com
For more options, visit http://groups.google.com/group/chinesemac
---
You received this message because you are subscribed to the Google Groups "Chinese Mac" group.
To unsubscribe from this group and stop receiving emails from it, send an email to chinesemac+...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/chinesemac/e9ef788a-6aa7-4420-9e80-4aeae2fcc8f2%40googlegroups.com.