TM language processing rules and recognized tokens
Translation memories (TMs) are complex databases that use a series of versatile elements catering for your translation needs.
TMs are based on language processing rules which include:
- Segmentation rules for the source text
- Recognized tokens (placeableas) to identify non-translatable content:
- Dates
- Times
- Numbers
- Measurements
- Acronyms and URLs
- Email addresses
- Alphanumeric strings
- Variables
- Currencies
The ability to generate and customize language processing rules makes the TM creation process very flexible and efficient.
TM backwards compatibility
Trados GroupShare 2020 CU3 and Trados Studio 2021 include major improvements to the language processing rules functionality. These improvements enable you to customize auto-localization formats for the following recognized tokens: dates, times, numbers, measurements and currency amounts.
The possibility to extend the formats used for identifying and auto-localizing recognized tokens is not supported in previous versions of Trados Studio and Trados GroupShare.
This causes the following compatibility limitations:
- Language processing rules created in Trados GroupShare 2020 CU3 are not compatible with TMs created in Trados Studio 2017 and 2019.
- TMs created in Trados GroupShare 2020 CU3 and later can be opened in Trados Studio 2017 and 2019 TMs, but you cannot edit the TM settings in these Trados Studio versions.
To avoid these limitations, we recommend using language processing rules created in Trados GroupShare 2020 CU3 and Trados Studio 2021, and later, only with the current and future releases.