A Tough Crowd
A few days ago, a blog post called One of the toughest job-interview questions ever started making the rounds. My biased summary: A fellow is interviewing for a job as a technical writer, and he is asked a very technical question:
Explain how you would develop a frequency-sorted list of the ten thousand most-used words in the English language.
He explained how he pushed back very hard on the requirements, then got into a long discussion about collisions and cuckoo hashes. He ended the post by suggesting that open-ended questions like this can be fruitful because they lead in unexpected directions and provoke lively discussion.
To my surprise, I have seen a lot of very negative comments about his post. Many have suggested that he was being difficult or obstructive to the process. The interviewer wanted him to give a technical answer to a technical question, not enter into a discussion about what "The ten thousand most-used words in the English language" actually means or whether it's feasible to collect a corpus of such words. I say I'm surprised because to me the whole point of an interview question is to open conversation. As long as he ended up giving a technical answer--which he did--what is the problem here?
What is worse to me is this: Let's say you decide you don't like this guy based on his responses. Okay, so what do you do with someone who just answers the technical question without pushing back on the requirements? Do you give them a pass? Think carefully before you answer. If you pass the person who gives a technical answer without pushing back, what you are doing is giving interviewees a huge disincentive to talk to you. If they do anything besides giving you the most literal, minimum answer to your question, they have nothing to gain and everything to lose.
This dynamic is common in interviews. Most interview processes start by selecting people who the interviewers think are qualified to do the job. They have the credentials and/or work experience required. The interviewers may not admit it to themselves, but the interviews are for weeding people out, not for identifying the best candidate. The difference is important, because if the process is about weeding people out, the candidate's incentive is to avoid doing wrong rather than to display aptitude. Interviewees are just trying to survive. Pushing back on requirements is not a survival technique. This is bad for you as an interviewer, because you end up with less information about which candidate is the most talented.
A final point.
Liking or disliking someone's "attitude" is very dangerous ground. What I believe happens is this: You decide whether you like or dislike someone in the blink of an eye. Then you cast everything they say in the light of your prejudices. If you have already decided that the candidate is "Just a technical writer," then his push-back is obviously the sign of a difficult personality who will be grit in the team's vaseline.
Or what do we do with someone who uses pushing back on the requirements to display their contempt for the question? Is being contemptuous of our interview process a bad sign? Why? Do we only want conformists who will never question us?
And what if, instead, we have decided that the candidate is one of the most charming, intelligent, and talented scientists in history, a Nobel prize winner? What if he shows you that your question was imprecise? In that case, maybe we will decide that pushing back on requirements is a sign that the question was a dumb one to begin with.
I find it interesting that a relatively unknown fellow gets slammed for his attitude, another fellow writes a roundly lauded post condemning technical questions, and the apocryphal story about Richard Feynman routinely draws praise.
Also interesting: My perception is that the above post is about How to interview people, not How to answer technical questions. I tried very hard to express no opinion about whether someone should respond as he did to a technical question where the interviewer explicitly tells you to make a bunch of assumptions and not think the problem through.
It's too small a sample to draw conclusions about communities and bias, but the early comments on news.ycombinator.com and programming.reddit.com show a sharply different take on the post. One discussion is about how to ask interview questions, the other is about how to answer them.
My recent work:
- jQuery Combinators, what else? A jQuery plugin for writing your own fluent, jQuery-like code.
(Spot a bug or a spelling mistake? This is a Github repo, fork it and send me a pull request!)