> Finally, an obvious point: evaluation of the performance of an individual software developer is one of the great unsolved problems of software engineering. I

> Finally, an obvious point: evaluation of the performance of an individual software developer is one of the great unsolved problems of software engineering. I don't think it's obvious at all. That an experienced engineer or manager can correctly evaluate a software engineer is one of the ground assumptions that runs through our industry and its thinkpieces. (Though it is borne out by the studies that, e.g., even Google has done). > If you think that's different at Google, re-evaluate: Google also runs one of the most famously capricious hiring programs in the industry. Despite constant rituals and genuflection towards data-driven decision making, Google continues to thrive based on its status as a premiere destination for new software developers, despite running a hiring process renowned for the quality of the people it has alienated. There is ample evidence of Google having scaled broken processes. This is a significant point: Google's hiring process is legendarily bad. While I would work for them (I think their ethics are solid, their engineering is excellent, their benefits suitable for adults, and reports of internal culture are generally good - a rare combination), the hoop jumping they ask you to go through is... quite amazing. I just don't have enough of a taste for doing an independent study of my upper division algorithms class to have aggressively tried to get hired there. It's also evidence, I think, that broken hiring processes do not inherently break a company - or at least a sufficiently profitable one.