The German verb ‘sichern’ gets different translations (borgen, back-uppen, zekeren, beveiligen/beschermen, …) into Dutch, depending on its meaning. For a CAT tool like CafeTran, there currently is no way to determine the meaning of a word. However, it can determine the context of a word at the surface level. When a word occurs in the context of certain other words (e.g. Schraubverbinding/Bolzen/Schraube, Daten/Datensatz/Diskette, Bergsteiger/Kletterer), this can give a cue about its meaning. So paying attention to a word’s context is a good way to narrow down the possible meanings of a word. This is where CafeTran’s Context-aware Auto-assembling can help you to be more productive.
Think of Context-aware Auto-assembling (AKA C-3A, see also C-3PO) as a kind of 3-dimensional glossary, where source and target each represent 1 dimension of the glossary and the third ‘dimension’ is represented by the context of a segment or project.
In the Context field of a source term (e.g. schließen) you enter words from the source language that require a specific target term (e.g. ‘Tür’ requires ‘sluiten‘, ‘Hahn’ requires ‘dichtdraaien’). Two simple glossary entries for ‘schließen’ would look like:
In the two examples below you can see that the correct translation of ‘schließen’ has been auto-assembled, based on the source words ‘Tür’ and ’Hahn’:
Let's have a look at the glossary file. Here, several target terms are combined in one row (record). They are separated by a semicolon. The Context field represents the third field of the glossary text file:
You can use regular expressions to combine context determiners:
NOTE: The pipe character ‘|’ is used here at the target-side to separate context specifiers (determiners).
Note also that the first line of the Context field is overqualified … The same effect is realised by [Mm]eldung|[Ww]arnung (which covers: Sicherheitsmeldungen, Meldungen, Hochwasserstandsmeldungen, Warnungen, Vorwarnung, Hochwasserwarnungen, Sicherheitswarnungen …).
Example 1: The source segment contains 'Fehlermeldungen', which is the plural of 'Fehlermeldung', which is present in the context field as a trigger for the target term 'weergeven'. In other words: since the source segment contains both 'Fehlermeldung' and 'Anzeigen', the target term 'weergeven' is inserted.
Example 2: The source segment contains 'beachten' and 'Anzeigen', so 'displays' is inserted.
Example 3: The source segment contains 'beachtet' and 'Anzeigen', so 'displays' is inserted.
Note that in cases where no context info has been supplied, CafeTran will take the first target term for auto-assembling. Please also note that in cases where no context specifier is found in the source segment, CafeTran will use the last context-specific target term for auto-assembling.
This makes sense, since it is very likely that in the remainder of the project word meanings will correspond with those in the first part of the project. If you nevertheless find that CafeTran has used a wrong target term for auto-assembling, you can easily override this via the right mouse button (context menu). Then, the new target term will be used for the rest of the translation project.
Another example:
German: entweder, oder
Dutch: hetzij, hetzij
Can be solved with this technique: …
