This article discusses the complexities behind measuring software productivity, suggesting that while inputs (like code written) and outputs (like features delivered) can be quantified, they don't adequately reflect the true productivity of software development. The author argues that software often does not directly produce value but acts as a vehicle to enable potential future value. Commenters argue that productivity is indeed measurable, disagreeing with the notion that it cannot be calculated or understood, pointing to the efficiency of problem-solving in coding as a valid metric.