Friday was the problem: intending to do one project, but another project steals the show. This is essentially what the last three years was like. As soon as I do "just one thing" on a project with momentum, it pulls me along for the rest of my time, leaving first tasks with little or no attention.

It's a tale of two inertias, with the project you should be doing having stalled inertia and the project you should ignore having moving inertia. Momentum is a good thing, regardless of where you point it, so the "just say no" approach isn't optimal. Stalling your ignore project so you can pay attention to the other isn't the way to go.

In track cycling there's a race called the Madison, where you switch riders after a number of laps in a longer race. The first rider has considerable momentum as she comes into the switch. The second rider has none. To keep as much of the momentum going, they grab each other's arms and do a "hand sling" where the first shoots the second rider forward. This works surprisingly well.

Get started on the project with momentum, doing your usual start of day review, which gets your "passive momentum" activated somewhat, then switch. The trick is to switch before the pull of the first project is to great. To help this, I usually make a mix tape for each project and start each day listening to the same songs. Listening to mix makes me feel like I'm in the momentum project. As I switch, the flow part of me doesn't even notice because of the power of the music.

This may all seem odd to someone that's unfamiliar with flow. The quality and productivity you get from maintaining a good coding flow is remarkably better than the usual event-driven fare. More on this later.