Get creative with your questions, not didactic. The question should be an exercise in thought, not one of rote learning. The point is to ask questions that are adaptive, as in the solution can get as deep as the interviewee/er is comfortable going. Esoterica belongs in the land of RTFM, unless the candidate’s solution leads there.

Unimaginative questions reveal unimaginative managers and coworkers. In practice, don’t give a brain teaser unless you authored it – and expect to find exceptions and alternative answers even if you did. No one has a lock on creativity and innovation.

Just one man’s humble opinion.

This post was mentioned on Twitter by blue64: In a #tech #interview should you be asking questions like “How many golf balls fit in a bus?” Interesting view here: http://bit.ly/4O8A1X...

There are several problems I see with these questions not already mentioned above:

1. They can be studied for, and therefore, used to game the system. Meaning you find out nothing about the person’s thought process, only their ability to pre-Google questions.
2. Even if you get a bad answer, that’s not sufficient to eliminate someone because you may just be missing their sweet spot.
3. You are often limited in time, these questions consume a great deal of time, and therefore, if you use only these questions, you are missing crucial knowledge about the candidate.
4. Even though you point out NOT to do this, these questions are still used by people as a “bozo filter” and it doesn’t work, period.

Three cases in point:
One and two, I hired two developers using these kind of questions. Both had impressive resumes, both answered the questions “well” and appeared to be good thinkers. Both developers turned out to be mediocre at best. One was always mired down in excess complexity to the point no one else could understand his code, the other was simply unproductive because he insisted on using different tools than the rest of the team and spent most of his time fooling around with them, rather than getting work done. Both were Smart, both were unable to Get Things Done. Had I spent more time on getting references of references, I might have discovered this.

Three, another developer I hired (someone I worked with previously and knew about his amazing work ethic) went through this process and was asked these questions by others. They all told me that I shouldn’t hire this person because they believed he wasn’t right for the job. I hired him anyway, and he was my #1 most productive engineer. Beat the socks off the other two guys above. And I still work with him today. He failed the question because it didn’t really expose the engineering abilities we needed. Again, the question failed to cover the necessary attributes: Are they Smart? Can they Get Things Done?

The reason this question is great is most people don’t know, and their gut reaction is usually wrong. So the question is a chance to see how they investigate a question. Using your criteria above (which are good), it measures “street smarts” and “ability to pivot.”

Please read the “why” I listed for the value of this question before dismissing it.