Dicksmoker said:
Oh for fuck's sake, VD writes perfect English (to the extent that that can even be measured)
Gonna go all Volourn on you. Bullshit. r00fles! Same with The Bitcher thing. Also, moran.
spectre said:
It is salvageable. It's a bit faster correcting even such machine translations then starting from scratch again, see my point about deadlines again. (Ok, I admit I'm just judging it on the fragment in the OP).
K, maybe not an editor here, but a senior translator-coordinator for a group of underlings.
This way you get consistent quality and decent speed, and imo it is THE way to localise stuff. Unfortunately, it costs it's value in money.
In my experience, it's the other way around. Correcting machine translations is a chore, and can take much longer than producing a translation from scratch. Because, well, it basically IS like translating from scratch, in a way, but you've got to trudge through badly written shizz in addition to that. Especially when you're dealing with a language that is not yet quite machine-translation-friendly (f.i. French > English would work pretty well, but Russkie > English - not really) - plus, getting the most out of machine translation requires the services of someone skilled in MT to calibrate the software - which is quite expensive in the first place, unless you're using Google Translate - and these guys don't come cheap.
Plus, when looking for editors for crappy MT, most people go around and ask: "Oh, hi, guize, we've got some stuff, could you give it a once over at your best rate?" (Neat mindtrick there - "best rate" actually means "the lowest you're willing to accept, bitch"). The hapless editor agrees, looks at the text, and commits ritual suicide.
Deadlines are not the problem, really - it's the workflow. First off, if the deadline's too tight, it means their project management is most probably off (especially considering that the stuff is being done internally, not by an outside vendor). Second, the translation itself should go through at least two stages of review, and I'd personally try to add a third layer when dealing with translators working OUT of their native language: first, as you said, translators reviewing and proofreading each other's work (that's mostly to avoid the "not-enough-time" problem you've brought up, although if a translator cannot do a good job given the deadline, he, technically speaking, shouldn't have taken the assignment in the first place, but we all know shit happens and you need the monies, etc.
), then general + technical QA that comes together with beta-testing the localization. For non-native translators, I'd go with another review by a native speaker, before sending the stuff out to QA.
Still, even all that can fail if the people responsible for the various steps in the process fail to deliver. And it's really, really easy to fail when working out of your native language, no matter how bad a dude you'd consider yourself to be. Keep in mind that I've got ubermenschlische standards when it comes to that, though; a hobby, you might say.