I do some work for a very large company which has thousands of documents translated every year and they do not use a single, online translation memory (what’s that?). Most freelance translators work online these days and if we could all use exactly the same translation memory it would be great.

At the moment however, for each order, this company’s agency forwards a tiny package (produced by the company itself) of the work to be translated and maybe a reference file or two. I am not even sure if the company’s preferred tool (Transit) supports an online memory. It would be good, malicious fun to run an analysis and see how much money they waste by having the same sentences translated over and over again.

I am guessing that this is because “it’s all too complicated” just like credit default swaps, no one knows what is actually going on, or how in this case translation memories actually work. All that I know is, that I occasionally recognise sentences which I have recently translated, sitting there like old friends waiting patiently for me to translate them again.

Since I have realised what is going on, I have started keeping everything from that customer and I manually include it, so that I then don’t have to waste my own brain power on repeat translations. I still get paid for it of course.

It’s still a shame though, because this company has lots of technical bits & pieces which are hard to name in English. It would be nice, when the terminology supplied can’t help, to be able search the memory of thousands of existing translations. Then I wouldn’t have to get researching in hyperspace and reinvent the wheel, which may end up a different shape to older ones.