In order to use a translation memory for analysis, pre-translation, or for the actual translation in the Memsource Editors, the TM must be assigned to a project first. Multiple TMs can be assigned to one project, and one TM can be assigned to multiple projects.
First, go to the main page for your job (you can get there by clicking on its name in the Projects section). From here, you will see a section labeled Translation Memories. Click the Create New button to create a new translation memory that will automatically be assigned to this project. Or, you can click Select and assign a previously created TM. Choose the languages you want and click Continue. You will then see a list of all the TMs you've previously created.
Selecting the TMs for a Project
There are several different options for each TM that you can use:
- This allows Linguists to read what's in the TM and must be checked if you want to use the TM. If you don't want to use a particular TM, then leave this box unchecked.
- Checking this box allows Linguists to update the segments in your TM that they feel could be improved upon. With this feature checked, any confirmed segments will be saved into the TM.
- Source Language
- This is where the source language of your saved TMs is displayed.
- Target Langues
- This section shows which target languages each TM has, regardless of which languages are in your current project.
- Penalty (%)
- Here you can set the penalty percentage for segments in your translation memory. See the Setting Penalties section below for more information.
There are a few important things to note:
- At least one TM must have the Write box checked. (The maximum is two "Write" TMs per language and workflow step in a project).
- There can be up to 10 TMs assigned to each project (per language and workflow step).
- Relevant TMs are displayed first. This is based on the Client, Domain, Subdomain and user rights of the Project Manager.
- A filter is available as shown in the picture above. You can use it to search for TMs by name, their ID number (on the far left side), or by the client's name.
- The first selected TM will be automatically selected for Write mode. You can change this if you select a different TM to be designated for Write mode.
Languages and Locales
As mentioned above, the source and target languages of your TM must be the same as the languages in your project. However, Memsource allows the selecting of TMs that have the reverse source-target combinations. For example, a TM that has English as the source and German as the target can be used for a project with German as the source and English as the target as shown below. These TMs will be read-only in the project.
Memsource also allows you to add a TM to a project with the same language but a different locale.
Generally, all languages with the same prefix can be added (EN, EN_GB, EN_UK, etc.). For example, if you create a TM with German and English as your languages, you can assign TMs that use both DE and DE_CH as well as EN_GB and EN_US (it also does not matter which is the source language and which is the target). However, keep in mind that all TUs will be stored as EN only with no distinction between US and GB. Therefore, using this TM for strictly GB or US projects might be inaccurate.
For certain languages, the code has a different prefix—for example, Norwegian has the code "NO" and Norwegian (Bokmål) has code "nb-NO”—these two languages cannot be mixed.
TM for Workflow Steps
When selecting TMs for projects with a workflow, you can decide whether the same translation memories should be selected for all of the project's workflow steps or whether each workflow step should have its own TM set up.
Note: You can select a TM for all workflow steps and all languages. Afterwards, you can assign one specific TM to one step and one language (if needed). However, this does not work in the reverse order; you cannot select a TM for a specific step and language, and afterwards select the TM for all steps and all languages.
If you are using proofreading on a regular basis, we recommended you:
- Create a TM for proofreaders only, where only reviewed translations will be stored (for example, TM_REV) and always assign it as Read for the translation step and as Write for the revision step.
- Create a TM for translation only (for example, TM_TRA) and assign it as Write for the translation step and READ in the revision step.
- You can even set a 2% penalty to TM_TRA so that the matches from TM_REV always have priority. This would mean that 101% matches from TM_TRA will be shown as 99% matches in T_REV. For more information on this, see Setting Penalties below.
- After a project is completed and all translation is reviewed and saved in TM_REV, you can then delete TM_TRA.
Setting Penalties for Translation Memories
If you assign multiple TMs to a project but want to prefer matches only from one (or a few selected TMs), you can set a penalty for any of the selected TMs, as shown below. These settings will impact the pre-translation, analysis, and the matches offered to the Linguists in the Memsource Editors.
An addition to applying penalties to an entire TM, you can also apply a 1% penalty to 101% (in-context) matches within a TM. This option is useful, for example, if you would like a Linguist to decide which 101% match from different TMs should be used in a translation. The penalty will mean that none of the in-context matches are inserted automatically into the target segment if the text is pre-translated.
In the Memsource Editors' CAT pane, the penalized matches will be displayed in a new order (a 101% match penalized by 2% will become a 99% match), with a little arrow indicating the penalization.
Articles related to Translation Memory: