Rob defends asking questions of PhD's after Ian decries them (is that the right use of the word?)

I have to say, I totally agree with Rob. If you are interviewing for a product group position (not research) as a senior software developer or architect, then I expect you to be able to code. I don't ask trick questions in interviews, I ask questions that I don't expect people to know the answer and then see them think.

I used to ask "write IntersectRect" all the time. It's a relatively simple problem, but people generally don't know the answer off the top of their head. The key is to see their thinking process, and ability to write code.

I remember interviewing an established PhD from another major software company, and when I asked him this question he seemed insulted. When I pushed for him to answer, he wasn't able to even code up the brute force approach to the problem. He may have been a fabulous theoretical compsci person, but he wouldn't have worked out well on a product team.

Now, to caveat this - we often divide the loop up looking to have each interviewer evaluate different things. In this case I was evaluating this person's coding skills. It's not to say that this person couldn't have worked out well at Microsoft, but rather that he did poorly for the skills *I* was looking for.