2 ms·
This is a bad interview question. It involves "rediscovering" a named algorithm on the fly, over the phone OR already knowing about it an acing it which is disi
by sitkack 13y ago
This is a bad interview question. It involves "rediscovering" a named algorithm on the fly, over the phone OR already knowing about it an acing it which is disingenuous on the part of the interviewee, or telling the truth that you know it and moving on to an equally bogus question. In the spectrum of questions, this one is near yes or no.
Good questions might start with, "how would you implement a system to ..."
or
"We have a system to solve VVV using UUU what issues good and bad do you see with this?"
- sliverstorm 13y agoIt could be a good interview question if they weren't looking for "the right answer", but the "we're out of time" after only the second go-round suggests it wasn't intended as the very-hard-question they want to watch you try to solve.
- rjzzleep 13y agoit's funny you say that. i once had an interview question from a bank saying how would you create a system that can find certain things in large xml documents faster than our competitors. I said: "I would use grep" so he asked, what if i also need to find attributes and some other stuff i can't remember. I said: "ah okay, in that case I would use a document store, like eXist" He asked: "but would it be faster than our competitors?" i start thinking how the hell would i know, and said: "i'd get something running, and then think about how we can optimize it afterwards" why the hell would i design a database system from scratch? needless to say he didn't really like what i had to say.
- walshemj 13y agoThe q woudl be how fast our are competitors? searching a huge document store for a given id say scale out using something like hadoop.
- yetanotherphd 13y agoYou should have told him you would run it on overclocked CPUs
- sitkack 13y agoI'd go get a job with the competition so I knew how faster their systems are. Also a ridiculous question.
- TLSaracen 13y agoI was under the impression that this question in particular was extremely common, and that most candidates already have heard of the solution. Because of this, I really hate this kind of question because you're getting an incredibly skewed view of the applicant's problem-solving skills. On the other hand, while design questions are important, there are other important qualities of a candidate that aren't best assessed by knowing how her or she would design a system (such as coding/problem solving/logic skills). I think the main point is interviewers should stop being lazy and come up with their own unique questions. For example, when I was interviewing at companies after college, I got the BST least common ancestor question at least 3 times. I'm sure the 3rd company had a much better impression of me than the first, even though I didn't get any better at problem solving. A simple question I sometimes ask is "through a football game, you can score (for simplicity's sake) some combination of touchdowns (7 points) and field goals (3 points). Write a function that takes a final score and outputs the number of possible ways a team could've reached that score." OP, best of luck on the rest of the interview process!