Brainbench perl test?
uri at stemsystems.com
Thu Sep 6 20:56:38 BST 2012
On 09/06/2012 02:34 PM, Abigail wrote:
> On Thu, Sep 06, 2012 at 03:55:36PM +0100, Peter Corlett wrote:
>> On 5 Sep 2012, at 17:35, Abigail wrote:
>>> No. Well, it filters out the wannabees. It doesn't recognize the serious
>>> coder. If, given the Fibonacci sequence, or a similar recursive formula,
>>> and your first instinct is to solve it with recursion or iteration, you
>>> aren't serious.
>> Isn't the *point* of this to be a simple test to quickly filter out the no-hopers? I'd hope it wasn't the *only* test.
> Well, I responded to uri who claimed it was the distinguish between wannabees
> and serious coders. But in my book, being able to calculate the Nth Fibonnaci
> number doesn't make you a serious coder. It just means you've past one hurdle,
> and I will have some followup questions. You haven't failed the interview,
> but you haven't shown that you're a serious coder yet.
maybe i overstepped in calling that a serious coder filter. i would
never just use that determining a skilled coder. it could be useful to
filter out the total losers. i speak to hiring managers all the time and
they give out similar tests just to filter out the losers.
i review perl tests for a partner agency. i have seen test results that
would make you blanch they were so bad. my favorite which i have seen
*multiple* times is for a basic merge sort. given a set of large sorted
files of lines, write an efficient merge sort. basic stuff and many do
ok with it. the dumbest ones read in all the files into a single array
and then call sort. then they wait for heat death. the second dumb
answer i see is a proper merge sort of only *two* input files into an
output file. then another merge sort of that output with another input
file until all the input files are done. faster than a full sort but it
reads in much of the data N times or so.
that is a much more real world problem than fib but the idea of
efficiency (which is key to this client) seems to get lost on too many
candidates. i have to give ratings on these tests and my partner decides
based on that whether to submit them to their client. as you would
expect, those answers get downgraded a lot.
More information about the london.pm