The point of giving FizzBuzz as a screening question was nothing more than that. A screen. It's just to know if the person can write any code at all. The initial poster of FizzBuzz didn't say he was looking for some of the crazy solutions people have been coming up with ever since, he was just looking to screen out the candidates who can't write code. I think FizzBuzz is a perfectly acceptable first question to ask someone. It's a question to know if you're not hiring the person. Not a question to know if you are. To know if you are going to hire the person there are much better questions to ask of course. And personally, I love all these FizzBuzz solutions that pop up everywhere. I think so far my favourite is Ruby interpreting Prolog' interpreting Lisp interpreting Lisp solving FizzBuzz : http://eigenclass.org/hiki/lisp-in-lisp-and-prolog-and-ruby On 6/6/07, Rick DeNatale <rick.denatale / gmail.com> wrote: > On 6/5/07, CHubas <CHubas7 / gmail.com> wrote: > > > > I really liked Rick's ideas > > > about what a Job Interview shall be, it is also for you to decide if > > > you want to work for the guys across the table. > > > > There are from jobs to jobs, but when one has to sell his own soul to > > the devil, at least has to impress him a little. > > Sorry if you're in the position of selling your soul, but if you are, > I'd suggest that you might not understand the particular devil you're > trying to impress and do the opposite. > > Last night, our local Agile meetup group had a talk by Ken Auer of > Role Model Software. He said some things which seem relevant to this > discussion. > > First, he said that he wasn't looking for 'clever' people when he > forms teams, he looks for people with a commitment to delivering > software and satisfying the client/customer. Too many programmers get > carried away with being clever and miss the goal, wasting resources > along the way. > > He told of a time some years ago when a new guy who'd been there a few > months called him over to show off what he'd been working on. He had > spent a week writing a binary search for a list being used in the > application. Ken asked why he'd done that, and he said because linear > search is slow. Ken then asked him how many elements were typically > in the list, and after thinking about it a few minutes the guy said > that he figured that there might be as many as TWENTY-FIVE! To top it > off, this was a Smalltalk project, and he'd been applying the C-level > knowledge he'd gained in his undergraduate data structures class, > duplicating things which were already implemented in the Smalltalk > collection classes. > > Personally I'm on the same wavelength as Ken. Better to have someone > with the basics, the right attitude towards the goals, and the > willingness and ability to help the team learn what's needed when > needed. > > Now this isn't to say that there are devils out there who would be > impressed, a lot of hiring managers aren't programmers themselves and > are easily impressed (or maybe I should say conned). I don't know how > many of those guys I've run into over my career who say "I'm not > technical" with a certain amount of pride in their tone, and even more > who thought they were technical and either never knew what that meant > or had forgotten it. Just between us, it seems to me that this whole > idea of using fizz buzz to screen applicants probably came from the > fevered mind of one of these latter types. > > Another thing Ken said that specifications on the whole are 15% > complete and 7% accurate. This is one of the main things which make > agile project management valuable. The specs WILL change, so be > prepared to change as you go. Build a little, find out where it's > wrong or where putting it in front of the user changes the perception > of the requirements, fix it, set the new short term goal and move on. > > It's interesting that the late recognition that the spec is ambiguous > shows this even at such a micro scale. > > Personally, being retired with some pension, I can choose which jobs I > want to work on, so I can try to find devil's with a compatible > perspective. What I look for are projects where I can first be > valuable and deliver whats needed as it's discovered, and second > educate myself along the way. > > If you hear of any such projects, drop me a line off-forum. <G> > > -- > Rick DeNatale > > My blog on Ruby > http://talklikeaduck.denhaven2.com/ > >