About the author: Jost Zetzsche is an ATA-certified English-to-German translator and a localization and translation consultant. He co-founded International Writers' Group on the Oregon coast and sends out a free, biweekly technical newsletter for translators (see www.internationalwriters.com/toolkit).
There
is a large variety of software development formats, and the most
logical distinction between different categories can be broken down as
follows:
binary files, i.e., files that cannot be opened and edited with a text editor, and
flat files, i.e., text-based files that can be opened and edited with a text editor.
Binary files include traditional formats such as .exe, .dll, or .ocx files. To translate these files, you will either
need a specific software localization tool that allows the direct
translation and necessary strings as well as further language-specific
development work and testing (for further information on processing
binary file formats, see my article in an earlier version of the Tool
Kit or in my ebook at www.internationalwriters.com/toolbox), OR
need to "decompile" these files with a development tool into
numerous individual flat files (and later "recompile" the binary file
again).
Many software developers view the decompile/recompile process as
outdated, but others don't and chances are that you will at some point
receive decompiled RC files to translate.
A section in an RC file could look like this:
CAPTION "Publish Web Pages" FONT 8, "MS Sans Serif" BEGIN LTEXT "Static",IDC_FILE_EXIST_STATIC,34,19,279,24 PUSHBUTTON "&No",IDC_NO_BUTTON,150,48,48,14 DEFPUSHBUTTON "&Yes",IDOK,24,48,48,14 PUSHBUTTON "Yes &All",IDC_YES_ALL_BUTTON,87,48,48,14 PUSHBUTTON "N&o All",IDC_NO_ALL_BUTTON,213,48,48,14 PUSHBUTTON "Cancel",IDCANCEL,276,48,48,14 END
The syntax is not too hard to understand from the translator's point of
view—most everything in quotation marks needs to be translated (but
not things like the font information or "Static") and the ampersand
sign marks that the following letter is a "hotkey" that you can access
with an Alt key combination in the software. Though it would be
possible to translate RC files in a text editor, it's not advisable
because a) you'll most likely overlook text that needs to be
translated, b) you'll tend to overwrite code where that should not
happen, and c) there is just no reason to not use your translation
memory for this. In fact, software files are rarely translated on their
own. Typically they are translated as a precursor to accompanying
documentation—documentation that will be using references to the
translated software over and over again—which is an ideal scenario
for the use of translation memory technology!
All of the above-mentioned localization tools as well as many
translation memory tools support the translation of RC files, including
Déjà Vu, SDLX, Star Transit, across, and TRADOS.
(By the way, while the translation of RC files in TRADOS before version
7 used to be rather tedious, it can now be done directly in the TagEditor which even offers certain RC-specific quality control features, including check for duplicate hotkeys.)
Here's a tip for the translation of RC files: Because RC files use the
quotation mark as a functional character, it requires a duplicated
quotation mark ("") for every linguistic quotation mark ("Click on
""Next"" to continue"). Some computer-assisted translation tools take
that into consideration—others may not. If you're not sure, you can
also use 'single' quotes which you do not need to double.
Many newer programming languages do not use a compiled format for their
resource files. Often this takes the form of XML-based formats (for
instance, the .resx example of .NET applications—I'll cover those in
a later newsletter) that are more or less supported by all major
translation environment tools, while in other instances still other
formats are used.
Java applications typically use the so-called Java Properties files (.properties).
The structure of Properties files is relatively simple: typically
everything that follows an equal sign needs to be translated,
everything that is preceded by a hash mark (#) is to be excluded, and
the translatable strings can contain some HTML-based tags.
Java Properties files are supported by most localization tools as well as SDLX, OmegaT, TRADOS, and Déjà Vu.
If you're not sure what format a certain software file is in (the
extensions are always a first indication of what the file type could
be, but they will often fail you with software files), open it in a
text editor and study the structure of the file. If translatables are
enclosed in quotation marks, try to process the file as an RC file (or,
if available, you can also test it with the C++ filter). If the
translatables are preceded by an equal sign, try to process them with
the Properties filter. All of these files are text-based, so this will
not damage the files and very often you will find that you "get lucky,"
even though the file at hand may not be one or the other.
Another emerging text-based software standard are GNU gettext PO and
POT files. These are the translatable language resource files used in
the free GNU gettext concept for translating software and
documentation. GNU gettext is the de-facto standard in many open source
projects, and it works with a large variety of programming languages.
PO files are typically translated or pretranslated files, whereas POT
files are the translatable templates.
Aside from the internal tools that gettext offers (see www.gnu.org/software/gettext), Déjà Vu seems to be the only translation environment tool that handles these files seamlessly.
Home
