Hi JF Delannoy ,

JF Delannoy said:

The report cannot be sorted as such.

You can sort the message list clicking on the header Message (or any other header), so you’ll have the inconsistencies grouped together:

JF Delannoy said:

As exported to Excel (or further to Excel, to Word), sorting is impossible because there are merged cells.

I don’t think it’s possible to export the message list, whether there are merged segments or not. If you can do it somehow, please let me know the way.

"Sentence was translated inconsistently - compared to segment 10."

but there is no direct way to change and propagage.

Suggestion: displaying both

and ensuring that a change will be recorded in MT and will be propagated.



With hundred of inconsistencies (because of lousy propagation), I can't handle this.

How do you ensure that the valid, unique, version of a segment in the MT propatages everywhere when you have opened the whole list of files precisely for this?

Run some function from the Translation Results pane?

Go to first occurrence and validate?

Inversely: if you inadvertently validate a wrong version, it substitutes for the good on in the MT, and there is no Undo!

Hi JF Delannoy ,

JF Delannoy said:

"Sentence was translated inconsistently - compared to segment 10."

but there is no direct way to change and propagage.

Suggestion: displaying both

To display both segments, I select the Source text in the Editor and then right click to filter by that content.

Now both (or several) segments will be filtered in Editor so I can edit them as needed.

JF Delannoy said:

How do you ensure that the valid, unique, version of a segment in the MT propatages everywhere when you have opened the whole list of files

This Auto-propagation feature doesn’t have anything to do with the translation memory (TM). It kicks in even without no TM. So the content confirmed in one segment gets auto-propagated to other segments (no TM involved, although usually the TM is updated after confirming a segment, but it’s better not to mix concepts). If there is a bug in Auto-propagation, as I think you said, I’d like to know about it. In File > Options > Editor > Auto-propagation you can decide how to auto-propagate (from the confirmed segment downwards or all segments, auto-propagate to already confirmed segments or not…).

Most of the time I use this settings:

I know that not all segments won’t be auto-propagated (only downwards, so I need to go to the first repeated segment) and already confirmed segments won’t be auto-propagated either (so I need to do something manual, basically deleting the translation of repeated segments leaving only the first one, then edit+confirm the first one to auto-propagate). Yes, it’s manual, but it works for me most of the time. If a Trados user is using these settings (by the way, the default settings) and believes that all segments are auto-propagated, he/she will get unexpected results. Other settings can be used as long you know what you are doing.

After the first round of QA has been done, I press F8 key to run Verify again until no actual errors are reported and only false positives are shown.

There are a few things I didn’t understand, so if there is still anything unclear, please let me know.

I have this settings and... it does not propagate.


I would expect it to jump to the next candidate and prompt.     Or, if this is sure ground, adapt automatically.



In his project, I have 9 files with company names, numbers acronums mentioned repetitively in 36 occasions.
I just passed all over for fish for non-adapted occurrences and change them (Ctrl T) manually.

I have this settings and... it does not propagate.


I would expect it to jump to the next candidate and prompt.     Or, if this is sure ground, adapt automatically.



In his project, I have 9 files with company names, numbers acronums mentioned repetitively in 36 occasions.
I just passed all over for fish for non-adapted occurrences and change them (Ctrl T) manually.