Trados7 adds multiple TUs 4 same SU even though "not allowed"! Thread poster: Hannah Robinson
|
|---|
On setting up my TM, I have not allowed multiple translations for the same source segments but Workbench seems to be storing some (not all) segments 2, 3 or 4 times, whether the target segment is different or not.
I have been revising the translation and saving the revised units, which should mean the old ones are updated no?
I don't know what is causing this or what to do to clean up the memory as I only want the final version of the translation stored there...
(I have also cl... See more On setting up my TM, I have not allowed multiple translations for the same source segments but Workbench seems to be storing some (not all) segments 2, 3 or 4 times, whether the target segment is different or not.
I have been revising the translation and saving the revised units, which should mean the old ones are updated no?
I don't know what is causing this or what to do to clean up the memory as I only want the final version of the translation stored there...
(I have also cleaned up the translation several times to proof it outside of Workbench - updating the TM in doing so - shouldn't the changed translations still replace the older versions though or am I doing something wrong?)
Project and attribute settings are the same for all the units...
Any suggestions? ▲ Collapse
| | | |
Vito Smolej
Germany
Local time: 00:17
Member (2004)
English to Slovenian
+ ...
SITE LOCALIZER | You can maintain TM or you can translate into it from source file | Jul 14, 2006 |
in eiter case, one of the two is out of sync. When synchronizing the two, make sure you do it correctly.
If you maintain TM, but forget to pretranslate, the source file gets outdated and will stay outdated - case I.
If you do pretranslate and happen to have "... to TM" checked, you synchronize TM to the old contents of the document instead of the other way around - case II.
Is there some additional way to do it wrong? Probably. I (painfully) remember th... See more in eiter case, one of the two is out of sync. When synchronizing the two, make sure you do it correctly.
If you maintain TM, but forget to pretranslate, the source file gets outdated and will stay outdated - case I.
If you do pretranslate and happen to have "... to TM" checked, you synchronize TM to the old contents of the document instead of the other way around - case II.
Is there some additional way to do it wrong? Probably. I (painfully) remember these two.
regards and hih
Vito
[Edited at 2006-07-14 13:18] ▲ Collapse
| | | |
| same problem | Jul 17, 2006 |
I have the same problem - multiple TUs with same Source Segment and same or different Target Segment (both possibilities).
But I cannot think of a way the duplicates had come into my TM although most units come into the TM by a regular CleanUp.
The only thing I can say definitely is: The higher the count of TUs in a memory, the higher the chance of duplicates.
The only way I know to kick the duplicates is to export all TUs to a file and then import it into a blank... See more I have the same problem - multiple TUs with same Source Segment and same or different Target Segment (both possibilities).
But I cannot think of a way the duplicates had come into my TM although most units come into the TM by a regular CleanUp.
The only thing I can say definitely is: The higher the count of TUs in a memory, the higher the chance of duplicates.
The only way I know to kick the duplicates is to export all TUs to a file and then import it into a blank TM with the "existing TU" option set to the option you want (mostly "keep most recent" or "merge"). The Reorganisation command do a very similar thing but you cannot determine which unit will be kept.
Does anyone have an idea how to securely avoid duplicates or kick them without exporting and importing?
thanks in advance ▲ Collapse
| | | |
| still not sure.. | Jul 18, 2006 |
Hi,
thanks for your replies! Vito, I'm not sure what you mean...I didn't pretranslate as it wasn't helpful to me on this occasion. So could that be the problem? But surely you don't have to pretranslate..?
My temporary solution has been to create a new memory and clean up the translation with "update TM" checked, which, of course, gave me a clean memory with each segment only once and also displayed them in the correct order.
However, I am concerned that as I cont... See more Hi,
thanks for your replies! Vito, I'm not sure what you mean...I didn't pretranslate as it wasn't helpful to me on this occasion. So could that be the problem? But surely you don't have to pretranslate..?
My temporary solution has been to create a new memory and clean up the translation with "update TM" checked, which, of course, gave me a clean memory with each segment only once and also displayed them in the correct order.
However, I am concerned that as I continue to translate, the same thing may happen..
I even carried out an experiment with a clean memory, translated a few segments and then altered them, some were updated and others were entered as separate entries into the TM - when I did exactly the same thing..!?
I'm sure the program is supposed to be able to handle this correctly, but can't see what I'm doing wrong and the userguide does not offer any other ideas either...? ▲ Collapse
| | |
|
|
|
| TM maintenance | Jul 18, 2006 |
I maintain Trados TMs in TMX format, using the TMX editor from Heartsome (shareware); there is Okapi freeware you can also use for this purpose, I believe.
By the way, isn't it possible that the TUs you think are the same actually differ in details (spaces, tags,...)?
Antonin
| | | |
Login to reply/comment 
Translation news related to SDL Trados
