The Computer Assisted Translation (CAT) pane displays segmented source text alongside matches from translation memories, term bases, and machine translations after pre-translation. Non-translatable matches are also displayed.
To see more information about a segment in the lower window, click on the segment in the CAT pane.
Hovering over information in the bottom window provides more detail.
If a target language Term has a Note or Usage information, it is tagged with an asterisk (*).
CAT results are displayed in the following order and are color coded:
Green - Best translation memory or non-translatable match, based on the score.
Yellow - Term base items (If there is a preferred term, it will be placed at top.)
Orange - Translation memory, non-translatable, or MTQE based on scores.
Blue - Machine translations (with no MTQE score).
Pink - Sub-segment TM (S).
101% - An in-context translation memory match.
A greater than a 100% match indicates the context also matches what is saved in the TM.
100% - An exact translation memory match.
78% - A fuzzy translation memory match (anything below 100%).
S - A subsegment match.
If a smaller part of the original text was previously translated as a short segment, the CAT pane will display it even though the match is lower then the threshold set in the Editor's Preference.
The downward arrow (↓) indicates TM penalization.
A project manager can set penalties for low quality TM matches in order to ensure they are reviewed. 100% matches, for example, may be shown as 95% matches.
TB indicates a suggested term from a term base.
TB in grey indicates a New term that has not yet been approved.
A term in red indicates a term that has been rejected and should not be used during translation.
When a term is selected, details are displayed at the bottom of the pane. Clicking on Edit source or Edit target, opens thepage.
A non-translatable (NT) is text that does not require translation. Memsource identifies these texts and suggests them as NT matches.
NT matches are AI-based; an algorithm working in the background identifies them based on specified criteria. NT segments usually contain characters, symbols, and words that do not need to be translated such as numbers, formulas, code, email addresses, currency, human and product names. As it is AI-based, is not possible to rule out any inconsistencies as the algorithm identifying the NTs is continuously being improved upon based on gathered data.
Depending on AI accuracy, the NT match is either 100% or 99% and are likely to be a segment not requiring translation but should still be reviewed by a linguist.
Display of the NT score in the Memsource Editor is optional and there are two settings:
NT with no score displayed with a white background in the score column and CAT pane
NT with score displayed as dash underlined green 100 or orange 99 in the score column and the CAT pane
To see match details:
Hover over the underlined score.
Click on the match in the CAT pane.
If enabled, the Memsource MTQE feature enables users to analyze a job and provides a percentage score for machine translation suggestions at the segment level before any post-editing is done.
100%: Perfect MT output - post-editing likely not required.
99%: Near-perfect MT output - minor post-editing required; likely formatting or punctuation issues.
75%: Good MT match, some human post-editing required.
No score: MTQE cannot confidently identify the quality. Output should be checked by a Linguist.
To set CAT Pane preferences, follow these steps:
From the Preferences.menu, select
Set the Minimum match rate for translation memory matches to be displayed and used for pre-translations.
Select the desired behavior after confirming a segment.
Enable/disable the auto-propagation of repetitions.
Admins and Project Managers can decide if the translations of repeated source text segments are automatically inserted into the corresponding target text segments. They can also choose whether Linguist users can control this setting.
For more information, see How to Be Secure with Memsource
Set Pre-translate options to control the behavior of pre-translating the next segment that the user jumps to after confirming a segment.
If machine translation is available for a project,will pre-translate segments with machine translation when a translation memory match above the minimum specified threshold is not available.
CAT preferences are saved for the editor.