One upside is that you're giving everybody the same test, so you don't have miscalibrated judgements of a candidate's ability, depending on if they used a compu

One upside is that you're giving everybody the same test, so you don't have miscalibrated judgements of a candidate's ability, depending on if they used a computer. It occurs to me now that another upside might be that it could have better predictive value. Maybe inability to deal with whiteboard code implies inability to deal with large code bases or complicated situations where you have to hold what the code's doing in your head. Or maybe that gets lost in the noise...