A personalized engine learned its terminology from your translation memories. It offers “e-mail” spelled with a hyphen and a patient “suffers symptoms”and a clinical trial “promotor” (in Spanish) is “sponsor” in English because these are the expressions in your TMs. Slate simply passed the source words untranslated to the target when source terminology is absent or “unknown” to your TMs.
When you need to override the engine’s learned vocabulary, such as using terms found in customer’s terminology glossary or when you experience unknown terms, you can prioritized your engine’s term offerings with in three terminology file options:
- Per engine terminology applies to one specific engine (language pair and subject).
- Per language pair terminology applies to all engines that share a specific language pair
- All engine terminology applies to all engines.
On-the-fly terminology updates
Finally you can update terminology while you work by using an AutoHotkey script. It reserves keystrokes that let you to select and add source and target language terms to your per-engine terminology file.
Slate Desktop gives you tools to mix translation units across various memories as one engine. You can indicate some translation memories are more important than others and the resulting engine will favor the terms and style in those important TMs.
If you consolidated many translation memories into a “big mama” TM, Slate Desktop has tools to split the “big mama” and curate the translation units back into discrete TM according to the properties stored in the TM.