DEMS Objects for Productivity Improvement (Top Texts)

The purpose of top texts is to help you reduce the manual translation workload and save you time and money.

When you run a job to create top texts, the system scans the source texts for one or more target languages, and then sorts them according to how often they occur in the objects relevant for translation. The source text with the highest level of repetition in the translation project becomes the first source text in the first top text object.

Translators call up the top text objects in the short text editor and create high quality proposals for the source texts that only require one translation in all contexts. After top text processing, you start an evaluation with automatic distribution so that the high quality proposals created by translators for source texts with a high level of repetition are inserted into translation lines in short text objects in the worklist and then saved.

There are two different ways of creating top texts:

• Top texts according to an object list

  You can run a job that scans the objects in a specific object list to create top text objects.

  You can restrict the job to object types that you have activated for translation into the specified target languages, or you can scan all objects in the object list irrespective of whether they are relevant for translation or not

• Top texts according to a worklist

  You can run a job that scans the objects in a worklist for one or more target languages to create top text objects.

The scenarios described under Examples below help you to decide which of the above methods is the best procedure.

This is how you integrate top texts into the translation workflow. Steps 4 to 8 can be repeated as often as you like during translation, even when other users are performing manual translation.

1. Set up the translation environment and define distribution parameters. For more information, see Setting Up the Translation Environment and Defining Distribution Parameters.

2. Create an object list that contains the objects you want to translate. For more information, see Creating Object Lists.

3. Generate a worklist and statistics by running an evaluation for the object list. For more information, see Creating Evaluations.

4. Create the users who will process the top texts as translators with a translator profile that enables them to create proposals with quality status A. For more information, see Maintaining Translators.

5. Create top texts.

6. Trigger processing of top texts. For more information, see Processing Top Texts.

7. After the translators have processed top texts, run another evaluation with distribution to regenerate the worklist and statistics that you created in step 3.

8. Compare the statistics that you generated in step 3 with the statistics you generated in step 7 for each target language to see how many lines were translated via distribution, thanks to the work on top texts.

The scenarios below illustrate that creating top texts according to an object list is only useful if you have just one target language, or several target languages that all need to translate exactly the same scope of object types, and collections in accordance with your translation graphs. You need to ensure that your object list only contains those objects that are relevant for translation. Otherwise, you may waste time during top text processing by creating proposals for source texts that are not relevant for translation.

In all other cases, creating top texts according to a worklist is a much safer option, because this job only scans the source texts that are actually relevant for translation for each target language.

The only exception is Scenario 5, which illustrates that top texts according to an object list can also be useful if you only want to create them for a subset of the objects in the current worklist.