3 ms·
Some very opinionated people in this thread. My two cents... I've had similar experiences to the guy who wrote the article. Something about interviews seems to
by lhnz 13y ago
Some very opinionated people in this thread. My two cents...
I've had similar experiences to the guy who wrote the article. Something about interviews seems to make my heart race and I can stumble on very simple code. This is in strong contrast to how I am in normal social situations where I am highly confident and my personality is often infectious and passionate. I think it's a case of high anxiety in authoritarian/test conditions. Those that imply their working conditions are similar to interview conditions make a good point, however if that is the case I would prefer somewhere less confrontational and more forgiving.
Generally I prefer proper timed tests [0] without anybody watching. This is how I normally work.
I'm also one of those people that has to constantly look things up; actually I have a broad range of experience but I often find myself delving deeply into things I have only cursory knowledge in - imagine a librarian that remembers the locations of the books in a library and the analogy to knowing high-level concepts that allow thinking on your feet when used together with the correct reference materials.
Of course I believe that it's better to be good at making do with limited information. Understanding things from scratch is what I have been doing all my life. Like a chameleon when I look at some code that I did not write, it becomes my syntax and I quickly understand the contents. However different organisations and cultures sometimes prefer different kinds of engineers: there is either a preference towards robustness or flexibility [1].
[0] http://codility.com http://codility.com
[1] A test I would excel on would be being passed code in an unfamiliar language, domain, library or ecosystem and being told to quickly get to grips with it and add features/fixes (aka. flexible knowledge: unfamiliar territory which requires me to pull together a broad range of skills.) This is opposed to being passed a piece of code in a familiar language, domain, library and ecosystem and being told to quickly and robustly add a feature or fix to it (aka. robust knowledge: familiar territory which requires only a very specific set of skills.)