I'm a big fan of coding in the interviews, and probably wouldn't work in a place that does not require so. What I don't like is being asked to code at home for several hours, perhaps never to hear from the company again, when the company is not willing to show any of their code.

@antonro Oh, I didn't understand that employers were doing that. That seems stupid and pointless. They would have no idea if the candidate was doing the coding or StackOverflow. I would not apply to those places. In what country do employers have you do this?

I'd say pretty much everywhere. Search for questions on "code challenge" or "coding task" on this site, you'll see.

@antonro OK I see - we do that level of screening first too - but with a multiple choice test which takes 20-30 min, after which we tell them their score and whether they have been advanced to the next step.

@antonro Personally I feel the complete opposite. I hate doing white-boarding or live-coding during an interview, I'd much rather spend a few hours coding the test at home with nobody looking over my shoulder then just discuss what I've done during the interview.

@SeanBurton The problem is that the interviewing company has to assume that you've written the code you presented. Surely you understand how easy it would be to get someone else to write it for you?

@SeanBurton Yes but they have to perform two steps to learn that in your method. If they make you write it during the interview, they've only wasted one step.

@JeffC but the first step is literally free, because it's not costing them any time if someone is coding at home.

@SeanBurton It's not free because someone has to look at it... and after looking it, what have you really learned? You've learned that someone (maybe the interviewee or maybe someone else) can code. You still have to proceed to the next step to learn whether it's the interviewee that can code.

@JeffC I'd very much prefer to code at home rather than in the broken artificial and anxiety-causing environment that white board coding interviews cause. Any interviewer worth their salt should know how to discuss the code challenge to determine if the candidate did it himself or not (i.e. what questions to ask). If they choose poor interviewers, then they're probably not qualified to be interviewing anyone in the first place... 1/2

@JeffC ...besides, coding interviews never replicate actual working conditions, and for a profession where your job is to think, it seems pretty dumb to place candidates in an artificial environment where they can't think properly to then turn around and claim they must be "bad" software developers. 2/2

@JeffC Yes, in your hypothetical situation you've wasted some time, but how often does it really happen that someone pays someone else to do a job's coding test for them? Maybe in other areas/countries it's more common, but in my experience It's going to be rare enough that it's not really an issue.

Its not really a 2 step method anyway, its 3, read CV, review test, interview. 2 of which takes 10 minutes, the last potentially a lot more. Now when you factor in many canidates, those for 10-20 minutes can really cut down the process.

@SeanBurton "but once they start asking questions about the code "you" wrote it will become pretty obvious to them if you didn't actually write it" Unless of course the person actually understands the code and is capable of answering questions about its structuring. Choosing to plagarise code doesn't necessarily mean a person is an incapable programmer, it just means they're highly immoral.

@Pharap, Sure, but if a person understands how to do the exercise, it would be easier to do the exercise themselves. It is certainly easier to explain the code that you wrote then to explain code that someone else wrote!

@Akavall What if they understand the language but don't understand how to do the excercise? Say for example a person is perfectly competant with a given language, but they've been asked to write a bubble sort and they know nothing about bubble sort. They could look at another implementation to understand it and then write their own version which ends up looking almost the same, or they could just take the implementation they understood and pass it off as their own. It's completely immoral, but it ends up saving the person time.

@Pharap So, you're basically saying that it's "immoral" to look at how others have done something to figure out how to solve a problem that has been solved before (e.g. sorting = re-inventing the wheel) during an interview, but if you're already hired, then it's perfectly "moral" to go to any coding website to do the same thing? Does that mean that the fact that they learned how to do sorting in college count as them being "immoral" since someone else already taught them how to do it and they may just be refreshing prior knowledge? I think you may be grasping at straws here.

@ray No, you misread what I said. I said it is immoral to plagarise the person's solution by submitting it to a company as if it were your own work. I did not say it was immoral to learn how to write an algorithm by disecting an existing implementation. The key turning point is where the hypothetical person chooses to skip writing their own implementation and submit someone else's code. (Submiting someone else's code and citing the origin would be less immoral, but that would probably hinder the person's chances of getting hired.) (Also they may not have learnt how to do sorting in college.)

@Pharap Yes, it's immoral to plagiarize, but can you really point to a piece of code that that's doing something that hasn't been done before? E.g., it could be argued that any bubble sort implementation counts as "plagiarism", even if you wrote your own version of it, b/c it's still the same algorithm; changing variable names wouldn't make it less of an issue, right? Trust me, I get your point, but it seems far fetched. Hypothetically, if you were to submit code to me for a problem you don't understand, you may be able to say what the code's doing, but not why it's doing it.

@ray I am saying that the hypothetical person does understand the problem and could write their own implementation but purposely chooses to submit someone else's implementation rather than spend time writing their own. If the person truly understands the algorithm and the nuances of the language, they'd understand why certain decisions had been made (e.g. why std::swap is used instead of a custom made swap).

@Pharap "I am saying that the hypothetical person does understand the problem and could write their own implementation" Ok, then what's the problem, exactly? Wasn't your concern that you could end up hiring someone that in reality would've been unable to solve the problem? (I'm not defending real plagiarism, but my prev. point on sorting algorithms stands.) You now have a person that, by your own account, is able to solve the problem, but used a shortcut. That's still an issue, yes, but it wasn't your original concern (i.e. whether they could or couldn't really solve the problem).