Details
-
New Feature
-
Resolution: Unresolved
-
None
-
None
-
Medium
-
-
Description
Certain patterns should be automatically recognized in segments and automatically replaced in the target, if it is configured like this for the current system/client/task. This will only be executed (and the configuration will point this out), if there is only one occurrence of a pattern type in the source segment at a time, because otherwise translate5 will not be able to decide, which is the first and which the second in the target of the segment.
Replacement means, it does not matter what the actual content of a pattern is, it only matters that it is the same type of pattern (for example a date in the same form of writing the date). Then the date in the TM (e. g. 10.02.2002) will be replaced with the date from the source segment (for example 24.12.2008) and the date will automatically be reconfigured to match the appropriate date format of the target language (e. g. 12/24/2008).
The match rate will be the same, as if the pattern in the TM already had been the correct one at the start.
The following patterns can be configured to be recognized:
- Numbers
- Date
- Time
- Acronyms
- Alphanumeric strings (containing numbers and letters in one "word")
- Units of measurement
Tags (and therefore also variables) will always be automatically recognized and replaced in the target.
The configuration for this feature will be located on language resource level.
The underlying idea how to solve this in the TM is to create two completely separate TMs on system level for every TM:
- one TM with the placeholders in the segments, where replacements are configured.
- If those settings are changed after TM creation, a warning message will be shown to the user that tells him that this will not effect segments, that have already been saved before his change.
- one TM with the original pattern.
- For querying matches and analysis the TM with the placeholders is used.
- The segment that is send to the TM to query for matches with all configured patterns replaced with placeholders.
- This will give the correct match rate for that match from the TM.
- In the front-end and in the pre-translation the actual
- For TMX export the TM without placeholders. This means for example, if placeholders are activated for numbers, a TMX export will still contain 2 segments for strings like those, despite they are saved as only one segmentn in the placeholder TM
-
- This is test 1.
- This is test 2.
- A TMX import must be parsed and the placeholders must be put into the TM and new duplicates that emerge through this procedure must be eliminated.