I'm going to differ from the other answers - I found this technique pretty common when talking to the HFT people. Lots of pet questions with extraordinarily spe

I'm going to differ from the other answers - I found this technique pretty common when talking to the HFT people. Lots of pet questions with extraordinarily specific answers expected. Example: implement the boost smart pointer on the board. I wrote something, started expounding on design choices and their implications, the guy cuts me off with "I said boost's implementation. Well, I didn't write the boost implementation, so that was a no go. Bullet dodged, imo. (actually bombed out during another question, some kind of loop detection algorithm. I got the answer, he claimed it was wrong, it just wasn't his way. I googled it afterwards to make sure I wasn't wrong.) There is a ton of high competition practice for the interview type stuff going on in that field, I believe. I kept far away from the field, so I don't have a lot of data points, and may be very wrong. But this is corroborated by the book Flash Boys - the interview techniques in that books were pretty much what I was subjected to.