Nothing grows on a timer: how a plant earns its next stage
When we first put grass into the world of The Long Watch, it was already there — placed, lit, saved, faithfully redrawn every time you returned. And it was completely frozen. Every blade sat at its full, mature size forever, like a photograph of a meadow rather than a meadow. This is the story of the small, stubborn decision that made it live: that a plant would never advance through its life on a timer, but only when the conditions it had genuinely lived through added up.
It would have been so much easier to use a clock. Plant a seed, wait a fixed number of seconds, promote it to the next stage, repeat. It would have looked almost right. We didn’t do it, and the reason why is most of what this post is about.
A life, not a countdown
A plant in The Long Watch moves through a handful of life stages — a seedling first, then a juvenile, then mature, then growing old at last. (A final stage, when the plant is truly gone, belongs to a different system entirely — what becomes of a dead plant is its own story.) The question that shaped everything is what tips a plant from one stage to the next.
The answer is not time. A plant does not count how long it has been alive and graduate on schedule. Instead it accumulates the conditions of its life — the soil it sits in, the moisture and warmth around it, the light reaching it — and that gathered experience is what carries it across each threshold. Growth is a running total of the world a plant has actually lived through, not a stopwatch ticking behind it.
Why a timer is a lie
A clock would have told a comforting lie: that every plant in the same place lives the same life. They don’t. Two seeds dropped on the same patch of ground fall through different seasons, different weather, different turns of light and shade. One sprouts into a kind spring and races ahead; the other lands as the warmth is leaving and barely moves for a while. On a timer, both would tick to maturity together, indifferent to everything around them. By making growth an accumulation, we let the difference between those two seeds be real.
So a favourable moment counts for a lot and a hard one counts for almost nothing. Whether a moment helps a plant grow starts with warmth: each species has a comfortable temperature band, and the further the world drifts outside it, the more growth slows or stalls. From there the soil beneath the plant and the light reaching it weigh in. A warm, well-fed, well-lit spell pushes a plant briskly up its stages; a cold or starved one all but pauses it. The plant isn’t aging — it’s responding.
Growth is the sum of the world a plant has lived through, not the time it has spent alive. A timer would only have told us how long; the world tells us how well.
Carrying the remainder
There’s a quiet detail here we’re fond of. When a plant gathers enough to cross into its next stage, it usually overshoots the line a little — it banks slightly more progress than the threshold needed. The easy thing would be to throw that overshoot away and start the next stage from zero. We carry it forward instead.
That choice matters more than it looks. If we reset on every threshold, the outcome would depend on how finely we happened to slice up time — promote in big steps and a plant loses more to rounding than if we promoted in small ones. Carrying the leftover keeps growth faithful no matter the step size, and lets a plant that has lived through a sudden burst of good fortune climb through more than one stage at once if it has truly earned it. The progression stays honest to the life the plant lived, not to the bookkeeping behind it.
The world remembers where each plant stands
Because growth is an accumulated history rather than a clock reading, we had to decide what happens when you close a world and come back to it. We store each plant’s gathered progress as part of the world, so a loaded meadow resumes exactly where it left off — no plant has to live its whole life over again from the seed just because you stepped away. Worlds you tended before this change load cleanly too: an old plant simply picks up from its last known stage and carries on.
The warmth a plant feels for growing, by the way, is the temperature of the ground it’s rooted in, shaped by the season turning overhead — not the air. That felt right for grass with its roots in the soil, and it’s why the same model will read the air instead for a tall canopy plant whose leaves live up in the wind, when we get there. The machinery is one tunable template, proven on a single species but written to welcome the rest rather than fitted only to grass.
Slowing it down without changing the answer
The hardest part wasn’t the philosophy — it was making it cheap enough to be true. Our test world holds around two thousand plants, and growth runs on the world’s own slow ecology cycle, a few times a second, well apart from the rate the screen redraws. Asking every one of those two thousand plants to read its soil and its light and advance, all in the same instant, was simply too much to do at once. The world would hitch.
So we spread the work out: each frame, only a small handful of plants take their step, and the rest wait their turn. The growing slows to a fraction of its nominal pace — but on a world that turns across in-game years, that slowness is completely invisible, and the world stays smooth, holding a comfortable frame rate with all the growing running live underneath it.
The thing we cared about most is that slowing it down did not change what grows. What’s fixed in this world isn’t the second-by-second pace of the work; it’s the sequence of growth cycles a plant lives through, in order. Spread that sequence across a thousand frames or run it all in one breath and a plant reaches the same stage either way. We checked it the only way worth trusting: that doing the growing gradually produces the very same result as doing it all at once, down to the last detail, and that the growth math behaves identically whether it’s measured in isolation or pumping away inside the running game. A world grown one way is byte-for-byte the world grown the other.
None of it leans on the wall clock. Growth is reproducible from a world’s seed and the conditions that seed gives rise to — the same way the same world always weathers and ages the same. (Keeping a living, surprising ecology and a perfectly repeatable save honest with each other is a whole craft of its own — how a forest is allowed to surprise us while a save never is.)
One rule, all the way down
We settled the design and built this in the back half of May, straight on top of the work that first made plants real things in the world rather than scenery. It looks like a small feature — grass that grows a little, slowly. But it set the rule we’d hold to for everything that came after. Decomposition advances by warmth and damp, not a decay timer. A creature wears out from age and hunger rather than running down a lifespan clock. Every link in the living world is a consequence of conditions, never a scheduled animation waiting for its cue.
That’s the whole point of tending a world instead of watching one. If things only happened on a timer, you’d be an audience. Because they happen when conditions add up, the world can surprise the person keeping watch over it — and the surprise is earned, the same way the grass earns its next stage. We started with the least important plant we had, and taught it to grow the honest way. Everything alive in The Long Watch grows like that now.
