Wouldn't it be good to know whether a segment to be translated is a heading, a warning, or a list item? The translation might differ depending on where the text is located in the document. For example, the German passage "Zylinder ausfahren" might be translated as "Extending the cylinder" in a heading, but as "Extend the cylinder" in an instruction.
Structure attributes enable you to provide your translators with this kind of information. Structure attributes provide additional information on your matches, indicating which part of the document structure a text comes from, e.g. "Heading", "Warning", or "List".
Structure matches are matches whose structure attribute corresponds to that of the segment to be translated. In other words, structure matches allow Across to verify that the text to be translated originates from the same document structure as the match that already exists in your translation memory.
Currently, you can use structure matches in tagged document formats, such as XML and HTML. Together with your translation partners, you could consider which structure attributes would be meaningful in your case and create these under "System Settings" > "Structure Attributes".
Go to the document settings templates and map the structure attributes to the respective XML or HTML elements.
Finally, customize the "Penalties" and "Pre-translation..." sections of your project settings templates, e.g. for your projects from your editorial or content management systems.
- Fewer translation errors due to missing structure information: Your translation partners can see at a glance that a text originates from a heading or warning and take this into consideration in the translation.
- Apart from the structure attribute information displayed in your matches, you can identify structure matches from the special icon they are labeled with in crossDesk.
- The results of your pre-translations can be improved significantly by systematically using structure matches.
- Structure matches are listed separately in the report and can significantly reduce your translation costs if the billing is arranged accordingly.
- Together with the context matches*, it is even possible
- to create structure-context matches – valuable matches
- that consider both the context and the document structure. Experience has shown that this approach saves your translation partners a lot of time for reviewing pre-translated texts.
- Translations variants for different document structures (e.g. "Heading" and "Instruction") are stored separately in your translation memory, thereby greatly reducing accidental overwriting or mistranslations:
- Structure attributes are single-value attributes, i.e. each translation unit can only be assigned one structure attribute. Thus, a translation comes either from a heading OR from an instruction.
- Depending on the language combination and frequency of use of structure attributes, this can result in "pseudo-duplicates", i.e. translation units that look identical at first glance, but that have different structure attributes.
- Of course you do not need to map every single XML or HTML element to a structure attribute. It is sufficient to use
- structure attributes where errors or misunderstandings can typically occur in the translation and where it would be helpful for your translation partners to know which structure level of your documents is concerned (see screenshot).
- As soon as you have defined the structure attributes and mapped them to the respective elements in the document settings templates, you should take a look at your project settings templates and modify them or create new ones if necessary.
*A context match will be found if the segment to be translated has the same context as the translation unit in your translation memory that was used for the pre-translation. For context matches, Across takes the segments directly before and after the segment to be translated into consideration.
Standard and custom training for new or advanced users at your desired date.