Having got working versions of the player movement and music in place, I asked my boy what he thought was the next most important thing to do, and offered some options:

1) map scrolling or screen switching
2) collision detection
3) time countdown

I think he was in favour of the screen switching, which was my initial instinct to be honest.

But as I thought about it, I realised that although a Pico 8 game is small, it's still a major undertaking for me. As soon as you go beyond one screen the possibility space of the game increases by at least an order of magnitude.

A better idea, surely, is to aim at the smallest/simplest thing that might possibly work and so this is my new idea:

• keep the action all on a single screen
• build a little maze with home at one end and school at the other
• have a countdown

This strikes me as a very obviously the best way forward. Scaling everything down makes so much sense:

• reduces the risk of not finishing
• allows end-to-end testing of race-against-time mechanic
• reframes the limited scope of the music as appropriate to the size of the effort
• makes meaningful exploration of other game dynamics more proximate
• although self-contained, the approach is still extensible: days of the week as levels
• etc

So that's great.

Next stop: collision.