A little over three decades here, and you're absolutely right. For every hour spent developing a cool algorithm, there are likely to be at least ten spent on lo

A little over three decades here, and you're absolutely right. For every hour spent developing a cool algorithm, there are likely to be at least ten spent on low-level coding details, hooking it up to its inputs and outputs elsewhere in the code, writing tests, debugging other people's tests, integrating with build systems, debugging other people's build breakage, monitoring and alarms, addressing code review nit-picks, rebasing on top of whatever my coworkers jammed in while I was being diligent, documentation, etc. Any non-trivial project will also be interrupted by meetings, project-wide direction changes, on-call shifts, and so on. If I get one uninterrupted morning/afternoon of working on "elegant solutions" in a month, that's a good month.