The reason we often end up working ridiculous hours doing HeroicProgramming is because SoftwareCostIsEasyToUnderestimate. Why am I always asked by people to write a "little app" that should only take a week? Why am I always being told by project managers that IfYouThinkYouCanAndYouTryVeryHard it will get done, usually at the end?
Software is really expensive. It costs a lot. This is frustrating to me that project after project stuff is promised that is really hard to deliver properly with the time allocated.
There are psychological considerations why underestimate is much more prevelant than overestimate. And these apply to all projects.
- you want to have approval for the project
- the sponsor(or management) want to demonstrate his prudence in budget management
- it helps to motivate people, work expands to fill up resources allocated
- maintain interest in stakeholders (see QuickWins)
I was just meditating on the CostOfLineOfCode. Maybe that little 1000 line app is more expensive than we think?