zlacker

This sort of stuff has parallels outside software - the book “how big things get done” is amazing, and one interesting thing it implies is that solar installations tend to go the best because they’re modular and lessons learned can be applied to the next phase of a project.

Whereas the worst overruns are nuclear power plants; which are either switched on and working, or completely useless and taking up space!

So my takeaways were: try to make estimates modular and work out how to carry learnings as your project goes on, and you’ll have an easier time hitting your estimates - and probably get a great reputation for delivering!