Fork me on GitHub
#jobs-discuss
<
2016-04-09
>
lmergen15:04:39

hey guys, a question

lmergen15:04:05

what was the most fun programming excercise while interviewing for a job?

lmergen15:04:50

i need inspiration, and i'm preferably looking for something open-ended

cky15:04:07

@lmergen: For my first job, the exercise was to be given a spec and a very buggy implementation of a word-counting program. You are timed on how long you take to fix all the bugs. Not sure if it counts as "open-ended" but the ability to debug under time pressure is very relevant to pretty much every programming job.

lmergen15:04:03

hmmm, i prefer to keep candidates comfortable rather than stressing them out -- a lot of them might already be pretty stressed because of the interview

lmergen15:04:20

i don't think if a candidate would fail, it would be representative

lmergen15:04:59

did you actually enjoy that, or did you enjoy it because it's different than usual?

donaldball16:04:44

I was never quite able to put this into practice, but when I was on the hiring side of the desk, my thought was to have the candidate implement a spec and then pair with them in adapting their program to new requirements

akjetma19:04:26

i had a lot a fun with this one (but i also had a lot of free time)

sveri19:04:29

@akjetma: Really? For an interview task? This is much more than I would do, timewise.

akjetma19:04:32

Yeah. I had fun working on it ¯\(ツ)

akjetma19:04:53

I agree though, it is kind of a ridiculously large question

akjetma19:04:00

‘question'

akjetma19:04:16

would unnecessarily filter out a lot of people

akjetma19:04:31

their interviewing process was pretty screwed up

sveri19:04:31

Yea totally, although it sounds fun simple_smile

lmergen19:04:50

i think a coding excercise shouldnt take longer than 1 hour, 2 hours max

lmergen19:04:19

if you cannot determine whether the candidate is qualified within that timeframe, you're not doing good enough a job as an interviewer

lmergen19:04:44

@donaldball: that sounds really interesting, btw! i might actually consider doing that, if the candidate's up to it

hiredman19:04:09

it varies a lot, shops give you are sort of take home problem, some having something they expect you to do during the interview

lmergen19:04:02

i think the take home problem is better, since the candidate can work on it in his own time, without stress... but i really like the pairing idea, since you get a far better impression about the candidate that way!

akjetma19:04:57

pairing during an interview can be weird/contrived if the interviewer is already completely familiar with the solution

lmergen19:04:16

yep, that would essentially be cheating

lmergen19:04:25

i was actually thinking more about the other way around

lmergen19:04:31

let the candidate choose the problem

lmergen19:04:36

and we solve it together

hiredman19:04:00

I find the take home stuff to be much more annoying, because like with all software problems are often under specified, or they are looking for something in a certain style, or whatever

lmergen19:04:14

it gives a very good impression about the candidate because of the type of problem he selects as well, what he thinks is relevant to show for the job

akjetma19:04:16

that’s a good idea imergen

lmergen19:04:42

im thinking about just working on a dojo or something

sveri19:04:29

@lmergen: But thats not fitting always too. The job I have now I had no idea what I will be doing in it during the interview, and not event three months after, which was the time when I just started to realize where I was heading too

lmergen19:04:50

thats true, but a good interviewer tailors the assignment towards the candidate

lmergen19:04:18

right now i have a candidate who's already very familiar with the type of problems we solve, so i'm trying to work with that

hiredman19:04:33

I just had a phone interview where we very quickly sketched out the implementation(using some online editor thing) of an app launcher thing (like for a phone), which I really enjoyed, because of the requirements it basically turned out to be building an LRU, and had opportunities to talk about the big o complexity of the app launching method, questions about how I would test it

hiredman20:04:07

so it started out seeming to be very, uh, I guess product focused, but ended up cs focused