Let's imagine following scenario:
I have to prepare a price estimation for the translation of a document. I don't want make a cost estimation for a translation from scratch so I have to create a translation memory. However at the moment I don't have the time to do the alignment in order to create a TM. This is thus what I do but it seems not to work:
- created a project.
- uploaded the source document of the previous version of a document (which is thus +/- similar to the document we have to translate).
- created a job and copied pasted all the source segments in the target segments in order to have a PM.
- put the job into completed.
- uploaded the document to be translated (the new version of the document).
No pre-translation or other matches as it should be.
The problem is from here : "created a job and copied pasted all the source segments in the target segments in order to have a PM. "
After this operation translation cells have been populated with the exact same text as the source. By default, the pre-translation process will not leverage this kind of hits because they are not relevant.
It is necessary to change this behavior by going into the project memory's search settings and tick the two options related as shown in the screenshot below.