this is all pretty good advice, especially the reviewing past [code] experiences and getting a sense for what it'd be like to work with the person.
i like staunch's "spend a day with the team" method a lot, though talking about strengths/weaknesses and role playing certain scenarios is also a good way to get a sense for potential clashes ("tests are for monkeys" or what not)
what i didn't understand was the "smart" part. i don't think it's possible to get a sense for whether someone is smart--especially technical smarts as opposed to, say, leadership or empathy smarts--simply by talking with someone. in fact, the whole concept of smart is exactly what pressure-ful, fanciful interview questions are trying to get at. i'm not even sure "smart" is a single or useful concept.
specific interview questions do try to determine technical skills or thinking process...but it a chat over coffee, a single, brief encounter, any different? in a good way?
i'm nitpicking here, but i felt that the solution to smarts--assuming we define smart to something useful--is kind of weak.
I suspect the "being smart" is in terms that the employer can relate to - do they think about things in the kind of depth that I like / appreciate / can work with.
From my personal experience, I've dealt with smart people and smart people - one kind knows / understands a lot of things in a broad field, the other kind knows a lot about things but is almost completely devoid of common sense. I'd prefer the first kind in any situation where I or my friends would have to deal with them over the medium to long term. I'm pretty sure this would be reflected in all areas worth measuring for a company too.
So trying to define smart as a concrete measure may be missing the point, as you will actually have to work with the person, and what you feel is what tends to form your opinion / bias and affects your future interaction in such a way that it generally tends to become a self-fulfilling prophecy.
i like staunch's "spend a day with the team" method a lot, though talking about strengths/weaknesses and role playing certain scenarios is also a good way to get a sense for potential clashes ("tests are for monkeys" or what not)
what i didn't understand was the "smart" part. i don't think it's possible to get a sense for whether someone is smart--especially technical smarts as opposed to, say, leadership or empathy smarts--simply by talking with someone. in fact, the whole concept of smart is exactly what pressure-ful, fanciful interview questions are trying to get at. i'm not even sure "smart" is a single or useful concept.
specific interview questions do try to determine technical skills or thinking process...but it a chat over coffee, a single, brief encounter, any different? in a good way?
i'm nitpicking here, but i felt that the solution to smarts--assuming we define smart to something useful--is kind of weak.