Conditional Text mid-sentence (using Lingo)

[doc_guy]

Hey all,

I've got a Flare project with a TON of conditional text. (We have three hardware models, so depending on the hardward model the client owns, the instructions are slightly different.)

When we created the Flare project, we used conditional text mid-sentence. What happens when I translate that text into multiple languages?

Take this sample example:

Press the {Condition1}Power{/Condition1}{Condition2}On{/Condition} button. A {condition1}blue star{/condition1}{green triangle} is displayed.

In both of these segments, the translation may affect other content outside of the condition tag. For example, the "a" in the second segment in other languages may be masculine or feminine, depending on the gender of the noun in the condition.

I searched the Lingo help high and low and coudln't find much information related to conditional text and how it is to be handled.

What do you do with conditional text in your projects? Do you need to do anything different when you are authoring to ensure your conditions are properly translated, and so they work properly in the final translated output?

Does it matter if you translate in Lingo or in an external tool? (We have some projects that will go to external translation, but our first project is being done by somebody using Lingo itself for translation.)

Any help you cna provide or insight from your projects would be greatly appreciated.

[whunter]

This has nothing to do with Lingo, it's just general advice. Back when I worked at a large company that handled a lot of translation, we were told that the smallest unit we could conditionalize is a full sentence. I think the reason for that is probably due to what you noticed below about different languages using different articles and whatnot, but also because the translation memory was stored per sentence.

Even conditionalizing per sentence was frowned upon a bit...some others advocated conditional paragraphs as the smallest unit, but I didn't really understand why. Maybe it was just easier for the translators to work on that way, when each paragraph made sense on its own.

[dmcleod]

In my projects this is handled more on the translators side. So on the first go through, the example you gave may be translated incorrectly with the wrong article gender, but when we receive that translation back, our content is all validated by someone internally who speaks the language. They will catch the mistake, and we will pass their feedback on to the translators, who will then plug the correction in to their translation memory and provide us with the corrected files.

So we proceed knowing there will be mistakes that must be corrected. Their frequency is low enough that the benefit of keeping everything as single sourced as possible outweighs the problems created for translation.