: How to handle bad source texts in technical translation? In technical translation (e.g. of manuals), the quality of the source text often is not very high. This concerns not only the language

Vote Up Vote Down

In the translation work I've seen for user-facing documentation, the translators stuck to the organization of the source but sometimes rephrased entire paragraphs, particularly if the source used idioms. This is particularly important if those translations need to be maintained over time as the source text is updated -- applying those updates to a translation that was rearranged will be both more expensive and more error-prone.

If there are problems in the source (you mentioned poor organization, for example), the best way to respond is to try to get those problems fixed in the source. This has the benefit of fixing it in at least two languages -- maybe more, if yours is not the only translation being done. My doc team occasionally receives requests for changes to make translation easier. We write with translation in mind to begin with, but nobody's perfect.

The general principle is to fix the problem as close to the source as possible. This is true whether you're talking about translation, formatting (single source, multiple outputs), or code.

(0)

Vote Up Vote Down

You mention manuals as an example. For these it may be down to the budget and the contract. For requirements documentation it is a different matter.

I have been involved in translations of requirements documents. These will contain clauses defining exactly how a system is to function once delivered. The document will almost certainly be referenced in a contract. These clauses will be explicitly tested by the customer and so must be retained and numbered identically in the translation.

It is necessary to translate as faithfully as possible, discuss the clauses with the original writer if at all possible.

(0)