My own 2 cents for interviewing, for engineers and most other vaguely technical positions but perhaps broadly applicable:
- Don't ask them to whiteboard, especially don't ask them to whiteboard some absurdity
- Don't put a panel of 5-10 people against a single candidate, but if you must absolutely (why?) do not do it on the first interview
- Don't ask time-wasting stupid questions about manhole covers, boiling bagels on mars, overlapping clock hands, or how many eggs fit into a phone booth
- Don't put someone into a socially awkward situation
- If you ask a knowledge question and the response is something like "I don't know, but I would google it" that should be acceptable if they are demonstrably capable of using what they find
- Do ask them to walk through their own projects, their code, their past work/contributions, etc
- Do try to effectively convey the corporate culture so the candidate can know if they are a good fit or not ahead of time
- Do try to accurately convey what their actual job duties and day-to-day work will be
It's just a variation on the typical logic/creative thinking questions that a lot of people and companies obsess over.
"Well they're a great candidate otherwise but they totally failed the boiling water on alternate planet question, so let's keep searching" - believe it or not, that's a real mentality out there. I don't get it and never have.
Maybe its just me but I find a lot of algorithmic interview questions are just as hit and miss as these obscure questions if you need to do it in an interview.
I had a phone interview with Google and was asked to come up with a variation of a soduko solver. Couldn't get it there and then, but half an hour later walking down the road, a really elegant solution came to me.
In context it sounds like a flippant reference to the 'creative thinking' type questions that were all the rage a few years ago. Ask someone an oddball question with the aim of having them talk through their process of solving a question they've never heard before because you just made it up.
As long as you make it clear that the actual number they come up with is irrelevant and you just want to see how they think (and you're also open to pragmatic answers like "why the hell would I be trying to calculate the weight of an Airbus A380 when I could just look it up on wikipedia?") I don't see why it's that bad a question.
I've taken recently a row of interviews. I've had my first whiteboard exercise and it wasn't as bad as I've expected to be. It actually proved to be very useful, as I could abstract away details that I otherwise had to deal with if I had to implement the solutions on a computer. I've also had those five interviewers on the other side of the desk! It was pretty funny for me, but I concede it may have been stressful for somebody else.
- Don't ask them to whiteboard, especially don't ask them to whiteboard some absurdity
- Don't put a panel of 5-10 people against a single candidate, but if you must absolutely (why?) do not do it on the first interview
- Don't ask time-wasting stupid questions about manhole covers, boiling bagels on mars, overlapping clock hands, or how many eggs fit into a phone booth
- Don't put someone into a socially awkward situation
- If you ask a knowledge question and the response is something like "I don't know, but I would google it" that should be acceptable if they are demonstrably capable of using what they find
- Do ask them to walk through their own projects, their code, their past work/contributions, etc
- Do try to effectively convey the corporate culture so the candidate can know if they are a good fit or not ahead of time
- Do try to accurately convey what their actual job duties and day-to-day work will be