FlowMission
The watch does the maths. You just go — six modes on one engine: an hour in a city you don't know, a race paced for you, or nothing at all except a clock that gets you home.
I like running. Or I would like to like running, which is not quite the same thing. This is the attempt to make an hour outdoors fun, a little exciting, and safe enough that somebody knows when to worry.
- I land at six and dinner is at eight. What is worth doing with the hour in between?
- I want to run 55 minutes. Tell me what to do right now, not what my average pace is.
- I know this city. Leave me alone — just make sure I get back.
- I'm 18 seconds down at 8 km. How much of that should I actually chase?
- Nobody knows where I am. Can someone see that I'm late, not just that I'm out?
What it does
One engine under five modes, and one idea: the watch works out what is possible and hands back a single instruction. Not a dashboard, not a route, not five true numbers to combine at threshold — one thing to do next, and a promise about how it ends. In a city that promise is that you will be back where you started before dinner. In a race it is the time you came for, or the best one still left.
- It does not turn round at halfway. Half the time is not half the distance — the way home is a straight line and the way out was not. The budget is computed once a second from your own pace, your own detour and the climb still to be regained, so a runner keeps exploring eleven minutes past the point a stopwatch would have turned them.
- It never chases lost time. A minute down at 8 km, a pace calculator asks for thirty seconds a kilometre faster over the two hardest kilometres of the race. That is how people end up walking at nine. The deficit gets spread and capped, and the rest is let go.
- It knows you are late, not just where you are. A dot on a map has no idea where you were going or when you meant to be back. This does, because the return point is where you pressed start — so the person at the other end can be told the difference.
- The far end alarms on silence, not on an alert. An app that promises to shout for help can only keep that promise while it is working — and a flat battery or a dead signal is exactly the case where the shout does not happen. So going quiet is the alarm. It is one file with no dependencies, on your own box: anybody trusted with knowing where somebody is should be able to read the whole of what they are running.
Which is why the beacon wants a way out that is not the phone: a safety net carried by the phone is missing on exactly the evenings it exists for. The open question is the grace period. Five minutes of lateness is nobody's emergency, but five minutes of silence would fire on the app's own backoff rather than on anything happening — so the offset comes with a rule attached. Once you are overdue the watch stops saving battery and asks every two minutes, because battery is cheaper than a missed alarm in the one state where that is true.
The walkthrough
Forty screens, the arithmetic underneath them, and every mission it can issue offline in a city it has never heard of. The last one is not drawn: it runs the same rules — the same detour factor, the same recovery caps, the same eight-sector rose, the same one wild card in four — so you can watch an hour happen.
Where it is
The modes run offline, building clean for five Garmin devices, with the figures in the walkthrough coming out of the app's own tests rather than out of a mock-up. The receiver is written and answers both the beacon and the place lookup, against OpenStreetMap. Trails now persist per city — off by default, coarse, thirty days — and what is left is an address to send to, not a build.
And it has never been worn. That is the next thing to do and the only one that matters: whether an hour of this is actually good is not a question a simulator can answer.
Is this worth building?
One press. No name, no address, no follow-up — it tells me whether to take this further or let it mature a while longer.
Noted. Thank you.