|
KnowHow: Troubleshooting incorrect extended characters in translation memory or target segments
Applies to:
TWB - 1.x, TWB - 2.00, TWB - 2.01, TWB - 2.20, TWB - 2.30, TWB - 2.x, TWB - 3, TWB - 5
OS:
WINDOWS - 95 (ANY), WINDOWS - 98 (ANY), WINDOWS 2000 - ANY (ANY), WINDOWS NT - 4.0 (ANY), WINDOWS NT SERVER - ANY (ANY)
SYMPTOMS
- In an existing translation memory, extended characters, for instance from Eastern European languages such as Czech, display fine. However, they are incorrectly transferred to the target segment in your document. In some cases, they may even display incorrectly in Translator's Workbench as well.
- In the translation unit, the incorrect extended characters are typically formatted with local font overrides, e.g. "This is an {\f23 ä} (a umlaut)."
- Font translation settings do not seem to have any impact on the problem, the characters are always incorrectly transferred to the target segment in your document.
EXPLANATION
The translation memory's font table is at least partly corrupted. More specifically, the codepage information for the font in question is incorrect.
In a translation memory's font table, fonts appear as follows:
{\f23 \fswiss\fprq2 \fcharset0 Arial CE;}
\fcharset denotes the codepage. In the example above, the codepage is the English one, with the effect that the font Arial CE displays any extended characters incorrectly.
This can happen when the Word documents being translated contain corrupt font tables due to a problem in Word. Incorrect information is then passed to the font table in the translation memory.
STEPS
You can fix the incorrect font table in the translation memory by exporting, externally modifying the exported file and re-importing it into a new translation memory. Follow these steps:
- Export the translation memory to a text file (.txt).
- Open the text file in any text editor, e.g. Notepad or WordPad. For bigger files, use a text editor that will accept large text files.
- Identify the fonts that should show extended characters but in fact use \fcharset0 (English codepage) in the <FontTable> section of the RTF preamble. For instance, if the font name "Arial CE" uses \fcharset0, correct it to \fcharset238 using the information below.
- Correct any other \fcharset values using the list below for reference.
- Create a new translation memory. Do not copy the translation memory setup from the existing translation memory, since this copies the font table as well. Instead, re-create the translation memory setup manually.
- Import the corrected text file into this new translation memory.
From now on, extended characters should display fine both in the translation memory and in your documents.
ADDITIONAL INFORMATION
To quickly find out whether incorrect extended characters are caused by a damaged font table in the translation memory, use the font preview feature in Workbench as follows:
- From the File menu, choose Properties.
- In the fonts list, hightlight the font in question.
If the text in the Example box reads "The quick brown fox", the font table contains the English codepage for this font. If, instead, you see several extended characters, the codepage information in the font table is correct.
REFERENCE
Font families and codepages:
Western fonts: \fcharset0
Baltic fonts: \fcharset186
CE fonts: \fcharset238
Cyr fonts: \fcharset204
Greek fonts: \fcharset161
Tur fonts: \fcharset162
Arabic fonts: \fcharset178
Hebrew fonts: \fcharset177
Shift JIS fonts: \fcharset128
Hangul fonts: \fcharset129
GB2312 fonts: \fcharset134
Chinese Big5 fonts: \fcharset136
Thai fonts: \fcharset222
STATUS:
Closed
Article ID: 780
Created: 07/06/2000 19:05:36
Last updated: 05/14/2001 11:37:58
|
