i got volunteered this evening for interview duty. Myself and assorted other suckersengineers get to do the first round talks with a batch of folks that we're trying to hire.
i hate interviews.
i don't mind the talking part, and i don't mind grilling folks about what they do or don't know, i just hate the whole pile of crap that goes with it. The fact is that few if anyone knows how to do an interview, particularly engineers. Inevitably they consist of either free form discussion about program design, compulsories or Click and Clack questions.
Remember how i said i hated interviews? i hate Click and Clack questions more.
For those not familiar with CarTalk on NPR, Tom and Ray always usually occasionally ask a puzzler. These are brain twister type puzzles that usually involve thinking about a problem in a nontraditional way. These are fun to muddle over because they're word problems with sufficient obfuscation to reduce the pool of folks who could win the $26 gift certificate.
The problem is that while seeing if someone can properly determine how to compare bowling pins using a 2×4 and a flock of northern European wood ducks, you still have no idea how well the person would handle detecting malformed URLs in a MySQL table.
i have a litany of questions that i ask candidates. The nice thing is that they not only tell me if the candidate knows his stuff, but also how he can present himself and complex ideas.
No cans of premium extra virgin olive oil are required.
Will i share this gauntlet of questions? No, not all of them. i will share one though.
"What's broken with our product? What would you most like to change?"
Bad answer: "Nothing. i like your site"
What you've just told me is that you either haven't used our product, or don't really care what happens with it. Like i said, bad answer.
The other nice thing about this litany of questions is that i can generally churn through folks pretty darn quickly. i usually can get an answer and opinion about someone within the first few minutes, and my opinion is more objective than "yeah, they seem nice enough."
which is probably the reason i got volunteered to do the interviews tonight.
*groan*
Maybe i'll start asking the candidates why you want to lean a big metal crowbar against a 1949 Ford truck that's being used to belt-drive a grist mill…
(oh, the title to this? It's what i answered someone at Microsoft who asked how i'd calculate the number of manhole covers in the US. i didn't get the job. It's ok, he couldn't answer why they weren't replicating pre-polled image caches across their data warehouses to reduce the out of network bandwidth charges. This was back in '99 so folks weren't thinking in those terms yet. It was a question i had after doing some research into their business model. i already had a job by the time they got around to doing second call. )
