> whiteboarding code by itself is not a great filter. It should be used in conjunction with a multi-faceted interviewing approach that involves testing fundamen

> whiteboarding code by itself is not a great filter. It should be used in conjunction with a multi-faceted interviewing approach that involves testing fundamentals, the ability to construct a relatively simple algorithm, the issues of working on a team and on a production code base and systems design. This I can heartily applaud. However, I take issue with this: > tests like the FizzBuzz test [1] are, in my experience, remarkably effective [...] I'd never heard of the FizzBuzz test before, just tried it. You may as well have asked me to sing the solution out loud. It was that bad. Paper turned out to be a completely unfamiliar medium for coding, despite years of whiteboarding solutions with colleagues. I instantly felt like I was using a totally different part of my brain. I practically lost the ability to write with a pen and significantly lost the ability to code. My heart rate rose substantially with fear of the unfamiliar. This tiny test took me ten minutes, and when I checked it later I found a syntax error. And that was in my favourite chair, at my desk, at home on a quiet sunny Sunday morning, working in a language I know inside out. Under pressure, with strangers, in an unfamiliar job interview room? I'd have just fallen apart completely. Back at the command-line? 30sec in a one-liner in a language I'm still learning. I will not be adopting this method to filter hires, having self-identified myself as a false negative. I am never out to embarrass someone in a job interview - I want them to show me their strengths! If they prefer it (they usually do), I'll give them a real computer and ask them to shine using that.