Re:Performance comparison: counting words in Python, Go, C++...
by hamburglar on Monday March 15, 2021 @08:58AM
(#26465543)
It seems almost disingenuous to me to miss the point of this type of interview question, because it is so obvious that the intention is “imagine a scenario where you need to do a high performance implementation of some very well-defined task, for which you don’t have a super fast utility function. Let’s use word frequency counting as an example and ignore the fact that such a problem is almost certainly already solved.” It’s implicit in the asking of the question that they aren’t actually in need of word counts on arbitrary bodies of text and need your help solving that problem. It’s a programming interview. Accept the premise and show them that you can program.
[ Parent
| on HN ]
Re:Performance comparison: counting words in Python, Go, C++...
by Veedrac on Monday March 15, 2021 @01:15PM
(#26468678)
It does irk me too. Being a programmer requires understanding the complexities hidden behind a task, and this brushes them all aside as if they don't matter.
There are cases where your task is nice, simple, and well-defined, but word counting is not one of them. It's better to give a more realistic task that allows people to think about the premises.
Re:Performance comparison: counting words in Python, Go, C++...
by hamburglar on Monday March 15, 2021 @01:42PM
(#26469017)
Being a programmer involves a lot of things that are difficult to test for all at once in 45 minutes. Be irked all you want, but it seems counterproductive to pretend what’s being asked is in anyway unclear or unreasonable when you’re given a pure programming question with a clear spec. You’re supposed to write the code. That’s all.
Re:Performance comparison: counting words in Python, Go, C++...
by Veedrac on Monday March 15, 2021 @01:46PM
(#26469068)
My problem isn't that the spec is unclear, it's that the spec is silly. The task makes no sense, and selects for misusing tools. This makes it a bad interview question as given.
Re:Performance comparison: counting words in Python, Go, C++...
by hamburglar on Monday March 15, 2021 @06:30PM
(#26470138)
I’d argue that the problem does make perfect sense and isn’t a “misuse of tools” if you manage to understand the goal of the interviewer, which is “watch this person write code for a solution to a well-specified but arbitrary (perhaps even silly) problem.”
Re:Performance comparison: counting words in Python, Go, C++...
by EvilEy3 on Monday March 15, 2021 @06:10PM
(#26469895)
This is litmus test to weed out people who pretend to be programmers. You assess complexities hidden behind the task in system design round.
Re:Performance comparison: counting words in Python, Go, C++...
by Veedrac on Monday March 15, 2021 @07:06PM
(#26470481)
I was more viewing this in context of the article, which goes on to micro-optimize a version with an IMO unreasonable set of assumptions.
If it's instead explicitly given that this is a quick litmus test, the choices of the first pass implementation don't really matter, and you'll talk about complexities after, then sure.
Re:Performance comparison: counting words in Python, Go, C++...
by hamburglar on Monday March 15, 2021 @09:32AM
(#26465875)
... otherwise the proper answer is probably “I would recognize I’m sitting in a room with someone who has likely already put enough thought into this task that they could be considered an expert on the solution and I would consult them on the best solution rather than reinventing the wheel,” which would almost certainly not get you hired.