Correlations Mechanism comprises automated processes that are running in the background during translation tasks "Analysis" phase to help decrease translation times as translations of segments which exactly match previous segments in the project can be anticipated and filled automatically. Its scope can be set up under local configurations of your project:
Correlations Mechanism Context Scopes
|None||The correlations mechanism is disabled.|
|Project||Translation tasks in "Analysis" phase will be compared to all the translation tasks in the project.|
|Labels (not implemented yet)||Translation tasks in "Analysis" phase will be compared to translation tasks sharing the same labels.|
|Resubmission||Translation tasks in "Analysis" phase will be compared to the translation tasks sharing the same Key/External Id.|
|Task||Only segments within the same translation task can be re-used to anticipate segments.|
The context scope should be chosen wisely to find the right balance between translation quality and time savings:
- If we anticipate the translation of a segment using another translation task that doesn’t share the same context, the translation might then be wrong.
- On the contrary, not using a segment that share the same context and which would most likely a relevant translation would lead to a profit loss.
Depending on the use of labels, having different content contexts in the same project, the use of resubmissions... different projects may need different context scopes. Finding a generic rule that would suit every project seems quite complex, thus a multi-level context scope configuration seems to be the best solution.
To find more information about general and local configurations, please visit this article: How to configure a project?