A friend gave me this link to a page of Joel's: 

http://www.joelonsoftware.com/articles/fog0000000073.html 

Pretty soon, my whole team was pinging it which would explain all those new referral links from Monday if you're reading this Joel.

The paper is a really important resource for interviewers. When interviewing people, I have followed these guidelines (probably before they were written) just on instinct. However as an interviewee, I've never had an interview that works this way. Mostly I get random technology-specific questions, and then I get random brain-teaser questions. I generally miss about half of these. For the technology-specific questions, I know a lot of technologies broadly, so that dings me on technical depth. (My understanding of Joel's paper is that this is discouraged, that aptitude is more important). As for the brain-teasers, my theory is that I tend to miss them because they are not real.   

Let me explain what I mean by "real." In the real world, problems don't come to us as disembodied exercises. The problems arrive in a dimensionally rich state. Most of the dimensions can be categorized as logical (are we thinking properly about this), and human (are we going to satisfy the market). Sometimes the human angle has more unknowns than the logical angle, such as aesthetic reactions by customers or time constraints from executives. Also in the human angle is a sort of recursive wording of the problem itself: the problem may be missing key data and that is a human problem for the bearer of the question (usually involving too much focus on a known area).

In an interview, brain-teasers arrive completely disembodied. There are suppositions which cannot be questioned. There is no human element in the role of the interviewer other than to watch the interviewee's behavior. The role of the interviewee is purely logical, and it would be improper to head any other direction. Can you imagine discussing the people who jump off the bridge with those marks on their heads from a human angle in an interview? No chance.  When I get a problem in real life, I take the human angle as most important, and work on the logic within that framework. Generally that results in getting things done. Generally I'm viewed as someone who "gets it."  Of course in an interview (the only time when I get the brain-teaser type questions) I’m left without the use of this “human element” strategy.

It is my theory that Mensa society members would have an easy time with the questions that I have missed during interviews. This makes sense, as these are the people who can get hired at software companies easily. They may not stay, because despite their mental powers they might not “get it,” but they’ll pass the brain-teaser with flying colors.

The take-homes for this situation are:

1) I should work on these brain teasers outside of an interview situation. If they’re obscuring the quality of what I can offer, like bad spelling, then fix it.

2) When managing a software project, assume that the human element is unfathomable by most people you’re working with. Get them to a position where they are technically challenged, and they won’t go off half-cocked looking for a market in the middle of the desert.

3) During interviews, make sure that I’ve demonstrated passion and aptitude, so that the brain-teaser is reduced in importance as well as time it takes in the interview. (This would be hard – ever try to manage an interview as the candidate?)    

One final funny story: A friend of mine at MS was listening to colleagues make fun of an intern who couldn't answer an interview question. To make a point my friend demanded that his colleagues all sort a doubly linked list right there at the lunch table. Let's just say many pieces of paper were used and lessons were learned.  

comment []12:12:37 PM