Under pressure and "in a strange environment" are two different things. You could put interview candidates in SCUBA gear and take them down 5 fathoms to see how they'd do under pressure, but that's also completely unrelated to predicting their actual job performance.
We don't ask trivia questions (what's the order of arguments to fputs, list all the overloads of foo in class bar) because knowing that (or anything else that's a 5 second lookup or context-sensitive help away) doesn't predict how someone will actually perform on the job.
Can any of us code as effectively on a whiteboard as we could in our favorite editor? I'd expect not, and if you were equally able on a whiteboard as in an editor, I'd wager that it's because you're not particularly good at either.
One thing I learned a while back at one interview (or rather, I just worked it out after experiencing it a few times) is that I find it much harder to think/code on a whiteboard. I don't know what it is, but it seems much easier for me to sit down with a pen and paper and work things out.
I mean, still coding without a computer - but it's something about standing up at a whiteboard. Maybe it's normally not so bad for me, but with the pressure of an interview, it's bad enough to give me problems.
We don't ask trivia questions (what's the order of arguments to fputs, list all the overloads of foo in class bar) because knowing that (or anything else that's a 5 second lookup or context-sensitive help away) doesn't predict how someone will actually perform on the job.
Can any of us code as effectively on a whiteboard as we could in our favorite editor? I'd expect not, and if you were equally able on a whiteboard as in an editor, I'd wager that it's because you're not particularly good at either.