CareerCup

Thanks for your reply!
Your reasons make sense.
(1) I don't think that they are so easy for a 30 min session..Is there any norm or is it entirely upto the interviewer..Though it may sound stupid but what is the expected time to solve a problem of avg difficulty level.
(2) Am I supposed to mention if I heard the problem before? Do you think that every interviewer gives new problem! After given a few interviews, I hardly found any original problem from the interviewers..All the problems are found to be discussed in the forums which I discovered after the interview!
(3) In one interview, they asked me to write the search function of a BST..Do they suppose it to be an unheard problem to the candidate!

(1) You will be compared to other candidates. So no, there' s no list that says "This is how long this problem should take." But if you ask enough candidates a question, you get a feel for what's fast and what's slow.

(2) Yes, you should mention if you've heard the problem before -- provided it's not a problem that you've obviously heard (see below).

(3) No, in this case you do not need to reveal that you've heard it before, because you obviously have. But if it's not obvious that you would have heard the problem before, you should reveal it.

Gayle,

I am curious. How does the interviewer realize (and be sure) that the candidate must have seen the problem before?

Has this even been a reason for rejection, in say the hiring committees you have sat on?

Also from the candidate's point of view. What about the cases where you saw it before and had solved it in say 5-10 minutes? Saying you have seen it before and solved it in 5-10 minutes might not sit well with the interviewer (he likely won't trust you). Isn't it better (and actually the true potrayal of your skills) to solve it again, as you did before?


Thanks.

"How does the interviewer realize (and be sure) that the candidate must have seen the problem before?"

A few indicators:
-- A candidate immediately knows how to solve a very difficult problem, but then struggles on an easier problem.
-- A candidate seems to be reciting a solution from memory.
-- A candidate is "struggling" but in a strange way. Their path doesn't match that of other candidates. They'll make sudden leaps of logic.

"Has this even been a reason for rejection, in say the hiring committees you have sat on?"

Yes. In fact, often candidates do worse BECAUSE they're trying to pretend that they haven't heard the problem before. And, candidates who admit that they've heard the problem before get "extra credit" in a sense because they're honest.

"What about the cases where you saw it before and had solved it in say 5-10 minutes? Saying you have seen it before and solved it in 5-10 minutes might not sit well with the interviewer (he likely won't trust you)."

Why would he not trust you?

"Isn't it better (and actually the true potrayal of your skills) to solve it again, as you did before?"

How do you "solve it as you did before"? If you know the solution, you'll solve it much more quickly. It's not a true portrayal of your skills at all.

Thanks a lot for your comments!
But the 3 points you mentioned seems to be subjective in nature and relying more on perception.
I would also like to have your comments on the following:
How is a problem solving exercise in an interview different from a one taken in an exam (in schools). Do we expect to get credit for being honest for a heard problem to answer.."Sorry..I have seen the problem..Can you change it!"...I am very curious about the interviewer's perspective on judgement of a candidate!

"How is a problem solving exercise in an interview different from a one taken in an exam (in schools)."

Because in an interview, they're trying to test your aptitude. They can't test your aptitude if you already know the question.

In school, they're trying to test your learning and knowledge.

"Do we expect to get credit for being honest for a heard problem to answer.."

Yes, most interviewers will.

""Sorry..I have seen the problem..Can you change it!"...I am very curious about the interviewer's perspective on judgement of a candidate!"

It's not up to you to request the problem be changed. You say, "I think I've heard that problem before. Would you still like me to solve it?"

Gayle,

1) People who can solve hard problems are known to stumble on the simplest of problems. There is no "monotonicity" regarding problem solving skills (i.e. if you can solve a hard problem, it does mean you should be able solve an easy one).

2) Just like you cannot judge if an interviewer is happy with you, the interviewer cannot be sure if the candidate is reciting a solution from memory.

3) Highly creative folks tend to make random, inexplicable leaps.

Over 5-6 interviews, yes, there might be strong evidence (but still circumstantial), but the possibility that a candidate might have heard a problem before should not be the _sole_ grounds for rejection. In fact, I would say, this is a failure of the interviewer/process rather than the interviewee (especially so, in a phone interview).

The 5-10 minutes thing, what I was also getting at was that being "obvious" is completely subjective. What is obvious to you might not be obvious to me or vice versa. Do you expect the interviewee to now gauge what is/is not obvious to the interviewer?

Most of the software industry is based on applying existing tools to current problems. There is very little innovation. I don't see why having seen a problem before (but having solved it honestly by yourself, of course) should count as a negative.

I meant...Do we expect to get credit for being honest for a heard problem to answer.."Sorry..I have seen the problem..Can you change it!"...in a school exam will be fetch me zero credit.
Now, by your logic knowledge/learning which are tested in school exams, do not portray skills/aptitude!
This argument will go to a different domain out of computer science.
At the end, I find every interview problem (barring a few from Google, etc.) is solved/discussed in some of the discussion forums..How can it be interviewee's fault when the interviewer is unable to put effort to pose new problems!

Thanks!

Yep. It is subjective. But that's how interviews are.

"Most of the software industry is based on applying existing tools to current problems. There is very little innovation."

Maybe... but that's not really relevant.

"I don't see why having seen a problem before (but having solved it honestly by yourself, of course) should count as a negative."

They're counting it as a negative because you're being dishonest.

I think you're misunderstanding the goal of interview questions. If you're asked, say, how you would find the median element in a set of sorted arrays, the interviewer doesn't care AT ALL in testing if you know how to find the median element in a set of sorted array. They're just using that question as a way of trying to figure out how good your problem solving skills are.

So if you have already heard the question before, that question will not tell them how good your problem solving skills are because you've already solved it (or been told the solution).

There are now two issues:
(1) You haven't given them any data with which to figure out your problem solving skills.
(2) You have been dishonest and tried to make your problem solving skills look better than they are.

"How can it be interviewee's fault when the interviewer is unable to put effort to pose new problems!"

It may be the interviewer's fault for asking questions that are too common, but it's your fault for being dishonest.

Specifically speaking, there is no agreement between the interviewer-interviewee that the interviewee need to disclose its list of heard problems...So what is the dishonesty here..The same thing is OK in a school exam but in an interview it is dishonest..Isn't it too subjective! I think most of the interview process is based on such subjectiveness/perception..Thanks for your comments and suggestions.

I don't see how it is dishonest, if you have solved it (by yourself and not seen the solution) before. What is the the harm in walking through your solution again? You can work through exactly as you did before, demonstrating what problem solving skills you already have.

Even if you finished a couple of minutes earlier, do you really think a few minutes difference is indicative of a significant gap in problem solving skills? (of course, you might say it depends on the problem, but we are talking about "non-obvious" problems here).

Note: the above is only in the case where you have solved it yourself before. I agree it is dishonest if you have only seen the answer and haven't solved it yourself.

You might get some extra points for the so called "honesty", but you are probably reducing your chances overall, as the interviewer might now pick a harder question from his list, making it likely that he is insufficiently calibrated on it, and would likely judge you with respect to himself (or the very strongest candidates who got to that question), rather than all the other candidates. Of course, if the interviewer has interviewed many people, this might not be a significant concern.

In any case, why should I hurt my chances trying to avoid perceived "dishonesty"? (After all, I _did solve_ the problem, only at a different point in time).

For instance, there recently was a problem about finding a string which is not a substring of a given string. I had never seen the problem before, but immediately got a suffix tree based solution. Why is it dishonest if an interviewer happens to ask me this, and I give the same solution? Why should I not take the chance to demonstrate my problem solving skills?

If the interviewer strongly suspects that the candidate might have heard the problem, they can always try to delve deeper, change the question a bit etc. This might even provide more evidence againt the "true dishonest" guys who haven't solved it themselves.

In short, I think interviewer should forget about trying to judge based on considerations of whether people have seen the question before, and concentrate on avoiding the situation altogether. I suppose Google does a decent job on that.

btw, the point of the comment of innovation was, there really are no new problems, so why should one expect that in interviews? You can always test problem solving skills by variations on existing themes or delving deeper. There is ample opporunity for that.

Basing your hiring decision on perceived "dishonesty" is a failing of the process, IMO.

Sorry for the long post, I do appreciate having this discussion with you.

btw, I am talking about the problems where it took me no more than 5 or 10 minutes to solve. Not the ones where I struggled for 30 minutes etc. I agree that if it took me a lot time, say 30+ minutes to solve a problem earlier, I will consider it dishonest to "solve" that in an interview in 5-10 minutes.

Thanks for your reply!
I am absolutely sure about my solutions since I have already practised the same problems before. For the 2nd problem, I discussed with him about the approach, and he told me to go forward with the coding. Then, I thought to improve the time complexity from n^2 to nlogn, which was approved by him and he told me to go forward with the typing.
I fail to understand the reason. If any specific requirement is there, then it can be cleared by a prelim session rather than going through the coding rounds!