On Sonja Tomakovićs Translators Blog, there is an interesting entry on quality assurance for translators.
In Germany, we usually associate that with the DIN norm for translators, DIN 2345. (There are two Austrian norms too). It’s possible not only to buy the norm – there’s an English translation too – but even to register and put little stickers on your work to say it’s DIN certified. This just defines process. I don’t seem to have mentioned it before – perhaps because it’s one of those religious war topics for translators, like Word versus WordPerfect, Trados versus Déjà Vu, or whether you should translate into a language that isn’t your mother tongue. Some translators believe that the very existence of quality control norms has a detrimental effect, but I’ve forgotten the argument.
Anyway, what is much more interesting is how we check our own and others’ work, which has as much to do with content as with process. DIN 2345 lists six points the translator is to check before delivering, stating that the extent to which each of these points is checked will depend on the purpose of the translation as agreed with the client. I find the DIN norm quite intresting to read.
The entry discusses an article in the ATA Chronicle, which unfortunately I haven’t seen (would anyone like my fax number?), referring to SAE standards:
bq. The article is by Don Sirena who works as a language translation manager for General Motors, and who has implemented the SAE J2450 quality assurance process for the translation department of GM.
bq. He points out that proofreading as we know it, i.e. simply going through a translation to observe errors, omissions or parts that we would do differently can lead to extremely high costs. Furthermore, the old picture of QA is that either you sacrifice quality for the sake of turnaround time and cost, or you have to live with higher costs when attaching importance to comprehensive QA processes.
The article apparently defines seven types of errors that can be checked for, each of which can be major or minor. This is reminiscent of the discussion of grading translation tests. Apparently time can be saved by doing this, since the final ‘proofreading’ is no longer necessary. (Proofreading and editing are often misused words, not necessarily here, but it’s necessary to be careful how they are used).
Not having read the article, I can’t see how this would save time.
I’ve recently changed my own self-revision process after reading Brian Mossop’s Editing and Revision for Translators. I’ve already mentioned the series Translation Practices Explained from St. Jerome Publishing. It’s a long treatment and very down-to-earth, and I took the author’s advice on what to read first for a self-revision workshop. Chapter 10, The Revision Parameters, sets out the types of error, in four classes and twelve subclasses. What is interesting is the later discussion, in Chapter 11, on how far to check these twelve matters, since there will scarcely be time to go through the text twelve times. Chapter 12, Revision Procedures, deals with the sequence of checking, and this helped me. It suggests that many translators check completeness and accuracy as they go along. In my case, I usually work in STAR Transit, and it’s easy for me to check every few sentences. The next step is now reading the translation alone, for which Mossop presents good arguments.
I’ve started using this procedure with longer texts. Previously I went through the whole in Transit, then did a comparative check either in Transit (onscreen or printed out) or Word, and this was lethally boring – the more boring a process is, the more slapdash I probably am. The reading through the whole translation in English is not the last stage, but it’s more fruitful than it used to be. I think my results have become more readable and probably more accurate. That may be just the novelty factor!
Brian Mossop has worked as a translator, reviser and trainer for the Canadian Government for 27 years and has taught translation at Toronto University for 22 years, and the book shows his experience.