Are you looking for a better way to organize your translation memories for reusing them in OmegaT?
We keep our translation memories in a centralized location on a network drive. This allows reusing the relevant translation memories in future projects efficiently, without having to copy them into each project manually. A centralized location also reduces the risk of some translation memories falling through the cracks. This article provides a more detailed description of how we do it:
The central location is Network drive:\Translation memories\Client tmx, with the following subfolders:
|auto||The folder used to insert 100% matches in a specific project.|
|Our TMs||Our translation memories from other projects.|
|Client’s translation memories||If available. Sometimes, we may penalize it if we believe the translations in our TMs are more recent.|
|Penalty-30||Translation memories from other projects in the same subject area. We use symbolic links to add them.|
Within the client folder, we also keep the glossary subfolder: Network drive:\Translation memories\Client glossary.
Using this central location in future projects
When a new project for the same client arrives, we have two options to configure the paths to the central location in OmegaT’s project:
- Find the previous project and copy its settings file, omegat.project, into the new project. OmegaT stores the paths to each project folder in this file. When a project is created through OmegaT, it contains default paths, pointing to the subfolders within the project folder. But when you copy it from the previous project, this file contains the paths you used there instead of the default ones.
- Open the project and change the paths manually in the Edit Project window.
Adding TMs to the central location
This is a manual process done regularly—for example, weekly.
- Open the folder with all your projects.
- Use the search function of whatever tool you prefer (we use Total Commander) to find files ending with “-omegat.tmx” created on or after the date of the last time you did this procedure. The files you will find are those translation memories exported from OmegaT projects that were created or changed since the last time.
- Copy those files to a separate folder.
- Manually move each of them to the respective client folder in the central location. If an older version of a file already exists, overwrite it.
- Rename the file as necessary. We name the files using this pattern: XX_project_name, where XX is a number that usually starts with 99 and goes down, with the latest TMs getting smaller numbers, so that they appear first in the search results in OmegaT.
It might be necessary to merge TMs. For example, if you have multiple TMs with duplicate entries, you may want to remove those duplicates. OmegaT is able not to display all duplicates in the Fuzzy Matches pane, but they still slow OmegaT down and display in the Text Search window. We merge TMXs with ApSIC Xbench, which is very efficient at this, but removes important information such as date and author.
For more information, refer to my first article about managing translation memories in an OmegaT-based translation workflow.