Luca Tutino wrote:
Discounting repetitions is always a dubious practice. For instance in my last EN-IT translation review I had segments including only the word "SCOPE", meaning "OSCILLOSCOPE" in some cases and "AMBITO" in others. However, this might be regarded as a relatively rare occurrence and you might be tempted to discount repetitions and exploit efficiently your expensive CAT tool. But if you use Trados, in my opinion you should not do it
. Read further if you want to know why.
The feature meant to avoid any double work is called auto-propagation. During the Review phase, auto-propagation:
- propagates changes and status of identical segments only if you change a segment
- does not
change the status of identical segments if you do not make any change in a segment.
This forces the reviewer to read practically all the repetitions.
You could think that the solution is simply to select a filter, but:
- if you select "only first occurrences", you do not see segments that do not have repetitions, and
- if you select "new translated content", you cannot be sure that wrong translations coming from earlier TMs are not included in what is hidden by the filter. (To be sure I had to find the definition of the "new translated content" filter in Trados documentation available online only
: "When this option is selected, segments with no match or a partial match that have been translated and confirmed are displayed." This means that if a translation comes from a 100% match of somebody's TM and is confirmed, you will not see it
using this filter - even if it is totally wrong...)
The result is that, if you want to be sure that you do not miss any error, in most cases, you will need to read every single repeated segment.
If the above is true, there is no basis for a Trados user to grant discounts for the review of files including repeated segments.
If you grant such a discount, you will inevitably end up working some time for free
, unless you want to compromise quality
NB I trust and sincerely hope that some member of the SDL community will read this post and prove me wrong.