[bencelaszlo]

Two things:

1. Ideas are cheap: not every one of them worth your time, or not that much time. Use your creative energy spikes to list and elaborate ideas. Estimation is a great weapon against scope creep, as you probably don't want to work on nice to haves for months. 2. Run to completion: if you decided to start a project, focus on it and continue until it is "done". Now, "done" is rare when we are talking about software, so the definition of "done" should include it is usable for someone for some purpose, have an executable binary, a build, a website, something.