CTRL+ALT+DIRT

CTRL+ALT+DIRT

The Great Developer-to-Farmer Migration

A Novel in Seven Harvests

2026 – 2032

“In the beginning, there was the command line.

Then someone said, “Have you tried turning it off

and planting tomatoes instead?”

— Ancient Stack Overflow Proverb, circa 2027

PROLOGUE

The Last Stand-Up Meeting

* * *

Nobody remembers exactly who said it first. Some claim it was a burned-out principal engineer at a FAANG company, staring at his fourteenth Jira ticket of the morning, who muttered, “You know what never has a sprint retrospective? Corn.” Others attribute it to a viral TikTok where a former DevOps engineer, still wearing noise-canceling headphones, was filmed successfully growing the most perfectly organized rows of lettuce anyone had ever seen.

Whatever the origin, by mid-2026, the Great Migration had begun. Developers—those pale, caffeinated architects of the digital world—started leaving Silicon Valley, Austin, Seattle, and every co-working space with a kombucha tap, and heading for the countryside. They traded their ergonomic standing desks for actual standing. In actual dirt. Under an actual sun that, as many of them would soon discover, did not come with a dark mode.

This is their story. It is a story of hope, of soil, of catastrophically over-engineered irrigation systems, and of one man’s eighteen-month quest to build a chicken coop using Kubernetes.

It is, above all, a love letter to everyone who ever thought: “Surely farming can’t be harder than debugging a race condition in a distributed system.”

They were wrong. But God, were they entertaining.

CHAPTER ONE

2026: git init farm

* * *

It started, as all great disasters do, with a Slack message.

“I’m done,” typed Marcus Chen, Senior Staff Engineer at CloudNova, into the #random channel at 3:47 AM on a Tuesday. “I’m buying a farm. I’m going to grow things. Real things. Things you can eat. Things that don’t require a YAML file.”

Forty-seven emoji reactions appeared within minutes. Thirty-two of them were the seedling emoji. Eleven were the tractor. Four were the saluting face. Nobody used the crying-laughing emoji, because nobody was laughing. They all understood. They had all stared into the void of a Kubernetes cluster at 2 AM and seen the void stare back, and the void was running out of memory.

Marcus wasn’t alone. That spring, something broke inside the developer community. Maybe it was the seventeenth JavaScript framework released that quarter (this one was called Turnip.js, which, in retrospect, was prophetic). Maybe it was the fact that GPT-7 could now write code faster than any human, but still couldn’t explain why it chose to name every variable “banana.” Maybe it was the realization that the average developer’s most meaningful relationship was with their compiler.

Whatever the cause, the exodus was real.

* * *

The first wave of developer-farmers made predictable mistakes. They approached agriculture the way they approached everything: with whiteboards, sprints, and an unreasonable faith in automation.

Jake Rodriguez, a former React developer from Portland, famously spent his entire first month building a “Soil Dashboard”—a real-time web application that monitored seventeen different soil metrics, displayed them in beautiful interactive charts, and sent push notifications to his phone every time the pH level shifted by 0.01 points. He had not, at this point, actually planted anything.

“I’m in the research phase,” Jake told his neighbor, Earl, a seventy-three-year-old farmer who had been growing soybeans since before the internet existed. Earl nodded politely, then went home and told his wife that the new neighbor was “one of them computer people” and that they should probably keep an extra eye on their property line.

Sarah Kim, a former backend engineer at a fintech startup, took a more systematic approach. She created a 47-page Notion document titled “Farm Architecture: A Microservices Approach to Agriculture.” The document proposed dividing her ten-acre property into independent “service zones,” each responsible for a single crop, communicating through a central “message queue” that was, essentially, a series of irrigation channels.

“The beauty of this approach,” Sarah explained in a Medium article that went massively viral, “is that if your tomato service goes down, your cucumber service remains unaffected.”

Earl, upon hearing this from his wife who had read the article aloud, responded: “That’s just… rows. She invented rows.”

* * *

The first great crisis of 2026 came in July, when the developer-farmers discovered a concept that no amount of Stack Overflow could prepare them for: weather.

Weather, they learned, was essentially a production outage that nobody could fix, that had no error logs, and that affected all environments simultaneously. There was no staging weather. There was no canary deployment of rain. When it hailed, it hailed on your tomatoes, your dashboard, your self-esteem, and your $3,000 drone that you had named “Jenkins” and sent up to “monitor crop health” but which was actually just taking aerial selfies of your farm for Instagram.

Marcus lost his entire first crop of arugula to an unexpected frost in late June. He sat on his porch, staring at the wilted leaves, and did the only thing he knew how to do: he opened a post-mortem document.

“Incident Report: Arugula Outage, June 28, 2026. Severity: Critical. Root Cause: Weather. Resolution: Unknown. Action Items: Investigate greenhouse (containerization for plants?). Time to Detection: 8 hours (was debugging unrelated irrigation API). Lessons Learned: Plants do not auto-scale.”

The document was shared in a Discord server called “Devs Who Farm” that had, by this point, accumulated 14,000 members. It received 2,300 upvotes and was later printed, framed, and hung in the lobby of Y Combinator, where it served as both an inspiration and a warning.

CHAPTER TWO

2027: The Year of the Merge Conflict

* * *

By early 2027, the developer-to-farmer pipeline was no longer a trickle. It was a flood. Real estate agents in rural Iowa, Montana, and Vermont reported that their most common buyer was a person under forty wearing a Patagonia vest, asking if the barn had good Wi-Fi, and wanting to know the latency to the nearest AWS region.

The rural communities handled this invasion with the grace and patience of people who had seen fads come and go. The Amish, in particular, were deeply amused. One Amish elder in Lancaster County was quoted in The New York Times as saying, “We have been living without technology for centuries. Now the technology people want to live without technology. We would like to point out that we do not offer tech support for horse plows.”

But the culture clash was real. Former developers brought their habits with them, and those habits were… specific.

Tom Wheeler, an ex-Google engineer who had purchased a dairy farm in Wisconsin, renamed all his cows after programming languages. The herd included Python (a slow but reliable milker), Rust (fast, never got sick, but was incredibly difficult to get into the barn), JavaScript (produced milk inconsistently and sometimes for no apparent reason), and COBOL (the oldest cow, whom everyone had forgotten about but who was somehow still producing).

“I tried naming one Haskell,” Tom told a reporter. “But she was purely functional and refused to produce any side effects. Including milk.”

* * *

The real problems began when the developers started trying to “optimize” traditional farming.

A collective of former Amazon engineers in Oregon created an automated harvesting system they called “Prime Harvest.” It used fourteen drones, a machine learning model trained on 50,000 images of ripe tomatoes, and a robotic arm salvaged from a decommissioned Tesla factory. The system cost $2.3 million to build.

On its first deployment, Prime Harvest correctly identified and picked exactly one tomato. It then, for reasons the engineers never fully understood, classified a garden gnome as “overripe” and attempted to harvest it, destroying both the gnome and the robotic arm in the process.

The incident was catalogued on GitHub as Issue #1. As of this writing, it remains open.

Meanwhile, Sarah Kim’s “microservices farm” had run into its own problems. Her independent crop zones, while theoretically elegant, had created what she described in a blog post as “the distributed systems problem of agriculture.” Her irrigation channels—the “message queue”—had developed leaks, resulting in her corn zone receiving 400% of its intended water volume while her herb garden had completely dried out.

“It’s basically a thundering herd problem,” she wrote, “except the herd is water and the services are thirsty, thirsty plants.”

Earl, reading this latest blog post aloud to his wife over breakfast, set down his coffee and said: “She’s got a busted pipe. Somebody tell her she’s got a busted pipe.”

* * *

The social dynamics of 2027 were perhaps the most entertaining part. Farmer’s markets, once sleepy weekend affairs, became battlegrounds of passive-aggressive one-upmanship between traditional farmers and the developer newcomers.

Traditional farmers brought their produce in wooden crates, hand-labeled with masking tape. Developer-farmers arrived with color-coded packaging, QR codes linking to “farm-to-table transparency dashboards,” and NFC-enabled price tags. One ex-Stripe engineer started accepting only cryptocurrency for his organic kale, which went about as well as you’d expect at a farmer’s market in rural Vermont where the median age was sixty-four.

“I just wanted to buy some tomatoes,” said longtime market customer Dorothy Hutchins, age 78. “The young man asked me if I wanted to pay with Ethereum. I told him I don’t speak Italian.”

By the end of 2027, an uneasy truce had formed. The developers admitted they didn’t know everything about farming. The traditional farmers admitted that the developers’ weather-prediction algorithms were, annoyingly, pretty accurate. And everyone agreed that Tom’s cow named JavaScript was a menace who needed to be kept in a separate paddock.

CHAPTER THREE

2028: Scaling the Harvest

* * *

The year 2028 was when the developer-farmers started to get good. Not great—that would take another year—but competent enough to stop being a punchline and start being a curiosity.

Marcus Chen’s farm had undergone a transformation. After the Great Arugula Incident of 2026, he had done something no developer had ever done before in the history of farming: he asked Earl for help. Earl, who had spent two years watching Marcus install soil sensors, crash drones, and once accidentally water his crops with sparkling water (“I grabbed the wrong hose,” Marcus claimed, unconvincingly), agreed.

The mentorship was a comedy of translation. Earl would say things like “the soil feels tired” and Marcus would nod and then spend three hours trying to figure out what metric corresponded to “tiredness” in his database. Earl would say “plant when the dogwood blooms” and Marcus would create a computer vision model to detect dogwood blooms from satellite imagery, which was technically impressive and also profoundly unnecessary since there was a dogwood tree visible from his kitchen window.

But slowly, improbably, it worked. Marcus learned to read the land not just through data but through experience. He learned that sometimes the best sensor was your own hand in the dirt. He learned that plants operated on their own schedule and that no amount of agile methodology could make a tomato ripen faster.

He also learned, to his enormous relief, that greenhouses were basically containers for plants and that his entire career in container orchestration had been, in a roundabout way, preparation for this exact moment.

* * *

The developer community, unable to resist its most fundamental instinct, had by this point created an entire ecosystem of farming software. There was FarmHub (“GitHub for farms”), CropCI (“continuous integration for agriculture”), and an app called Tiller that was basically Tinder but for finding compatible crop rotations. “Swipe right on soybeans,” the marketing copy read. “They’re great at fixing nitrogen and they’ll never ghost you.”

The most controversial product was AgileAcres, a project management tool that applied Scrum methodology to farming. Farmers would organize their work into two-week sprints. Each morning began with a stand-up meeting—literally standing, in the field, at 5 AM, which the developers had strong opinions about.

“The stand-up is supposed to be fifteen minutes,” complained one ex-developer, standing in the mud in Wisconsin at dawn. “This one involves chickens. The chickens don’t understand time-boxing.”

Sprint reviews were conducted at the farmer’s market. Retrospectives became legendary. One particularly honest retro, conducted by a farm in Vermont, produced the action item: “Stop treating the goats as stakeholders. They do not have actionable feedback. Their feedback is eating the sprint board.”

* * *

But 2028 also brought the first real crisis of identity for the movement. A Wall Street Journal article titled “The Great Reaping: Have Tech’s Farmers Actually Learned Anything?” sparked a fierce debate. The article pointed out that despite all the dashboards, drones, and disruption, the developer-farmers were producing, on average, 30% less per acre than their traditional counterparts.

The developer community responded the way it always did: with a hackathon. The “FarmHack 2028” event, held in a barn in Iowa, brought together 500 developer-farmers for a weekend of intensive problem-solving. Teams built everything from AI-powered pest detection systems to a blockchain-based crop insurance platform.

The winning project was, improbably, a simple mobile app that connected new developer-farmers with experienced traditional farmers for mentorship. It was called “git blame,” and its tagline was “Because sometimes you just need someone to tell you you’re doing it wrong.”

Earl was its first mentor. He charged nothing. He did, however, insist that all mentorship sessions begin with the phrase “Now, I don’t know nothing about computers, but…” regardless of whether computers were involved.

CHAPTER FOUR

2029: The Pivot Year

* * *

Every startup has a pivot. The Great Developer-to-Farmer Migration had its pivot in 2029, and it came from the most unlikely source: the developers’ children.

By 2029, the first generation of kids raised on developer-farms had entered their teenage years. These were children who had grown up watching their parents argue about TypeScript over breakfast and compost ratios over dinner. They were fluent in both Python and plant biology. They could deploy a web app and transplant a seedling with equal competence. They were, in a word, terrifying.

Fifteen-year-old Maya Chen—Marcus’s daughter—became the poster child of this generation when she gave a TED talk titled “The Full Stack Farmer: Why My Generation Will Feed the World.” The talk went viral, racking up 40 million views. In it, Maya argued that the fusion of technology and traditional farming wasn’t a joke—it was the future.

“My dad spent two years trying to automate everything,” she said, to laughter from the audience. “Then Earl taught him to actually farm. The magic isn’t in replacing one with the other. It’s in combining them. My dad’s soil sensors saved Earl’s crop last summer. Earl’s advice about companion planting doubled my dad’s tomato yield. The future of farming isn’t about code or dirt. It’s about code AND dirt.”

The audience gave her a standing ovation. Earl, watching from his living room in Iowa, turned to his wife and said, “Smart kid. Takes after me.”

* * *

The pivot manifested in practical ways. Developer-farms stopped trying to replace traditional methods and started enhancing them. The drones stopped trying to harvest and started doing useful things like aerial crop surveys and targeted pest control. The sensors stopped generating useless dashboards and started providing actionable alerts: “Your north field moisture is 15% below optimal. Earl says water it before Thursday.”

Tom’s dairy farm became a model of this new approach. His cows were still named after programming languages (the herd had grown to include Go, a sturdy and efficient milker; Kotlin, a newer cow who did everything Java did but with less ceremony; and PHP, whom everyone made fun of but who was somehow still the most productive member of the herd). But now the technology served the cows rather than the other way around.

Automated health monitoring caught early signs of illness. Data-driven feeding schedules optimized nutrition. And a simple app let Tom track each cow’s output with the same ease he once tracked API metrics. The farm’s milk production increased by 40%, and Tom became the unlikely star of a Wisconsin tourism campaign: “Come for the cheese. Stay because a former Google engineer named his cow after your favorite programming language.”

* * *

The cultural impact was unmistakable. In 2029, “farmer” overtook “engineer” as the most popular career aspiration among computer science graduates, according to a survey that made every tech CEO in America choke on their cold brew simultaneously. Venture capital firms, sniffing opportunity with their usual subtlety, began pouring money into “AgriTech 2.0” startups.

The most funded startup of 2029 was a company called Rootstock, which promised to be “the Salesforce of soil.” Its pitch deck contained the phrase “disrupting the dirt” on seven separate slides. It raised $400 million in its Series A. It had not, at the time of funding, grown a single plant. The developer-farmers, having seen this movie before, responded with a collective eye-roll so powerful it was briefly registered as seismic activity in three states.

“I left tech to get away from this,” Sarah Kim posted on the Devs Who Farm Discord, which now had 890,000 members. “If I see one more VC in designer overalls asking me about my ‘total addressable market,’ I’m going to feed them to the goats.”

The goats, as always, had no comment. But they did eat a visiting investor’s leather briefcase, which was widely considered a form of commentary.

CHAPTER FIVE

2030: The Great Harvest

* * *

In 2030, something remarkable happened: the developer-farmers started outperforming.

Not everywhere. Not all of them. But across enough farms, in enough states, with enough consistency that the data was undeniable. A study by UC Davis—co-authored by an agricultural scientist and a former Netflix engineer who had pivoted to studying crop yields—found that farms using the developer-farmer hybrid approach were producing 22% more per acre than the national average while using 35% less water.

The secret, the study found, was not any single technology. It was the mindset. Developer-farmers treated their farms like complex systems. They monitored, measured, iterated, and—crucially—they had finally learned to combine their data with the hard-won wisdom of traditional farmers. The sensors told them what was happening. The Earls of the world told them what it meant.

Marcus Chen’s farm was featured on the cover of both Wired and Farming Today in the same month—a first in the history of magazine publishing. The Wired headline read: “The Farmer Who Speaks Fluent Bash.” The Farming Today headline read: “Local Man Finally Grows Decent Arugula.” Marcus framed both.

* * *

The year also saw the rise of what the media called “Farm Stack”—a standardized technology toolkit that had emerged organically from the developer-farming community. It included open-source irrigation controllers, shared weather prediction models, community soil databases, and a particularly beloved app called “Stack Overground” where farmers could ask questions like “Why are my tomatoes splitting?” and receive answers that were, for once, not prefaced with “This question has been marked as duplicate.”

The developer-farmers had also, somewhat accidentally, solved the loneliness problem that had plagued rural communities for decades. The Devs Who Farm Discord had spawned local chapters, meetups, and—in a development that surprised absolutely no one—a podcast called “Ship It (The Compost)” that had 2 million subscribers.

Community barn raisings became community “barn-a-thons,” where developer-farmers would spend a weekend building structures for new arrivals. These events featured the same energy as hackathons—the same sleep deprivation, the same pizza consumption, the same moment at 3 AM where someone would suggest an insane idea that somehow worked—except instead of building an app, they were building a chicken coop.

“This is just like a hackathon,” observed one participant, hammering a nail at 2 AM. “Except if my code crashes, nobody dies. But if this roof collapses, chickens die. So the stakes are actually higher. Is this what real engineering feels like?”

* * *

Jake Rodriguez, the former React developer who had spent his first month building a soil dashboard instead of planting anything, had undergone perhaps the most dramatic transformation of anyone. By 2030, he had become one of the most respected organic farmers in Oregon. He still maintained his soil dashboard—it was genuinely useful now—but he had also learned to trust his instincts.

“The dashboard tells me the nitrogen levels,” he told a visiting journalist. “But I can also just… look at the plants. If they’re yellow, they need nitrogen. I know this sounds revolutionary, but apparently farmers have been doing this for ten thousand years without a single React component.”

He paused, then added: “Don’t tell my old engineering manager I said that. He’ll want a retro.”

CHAPTER SIX

2031: Legacy Code and Legacy Crops

* * *

The year 2031 was the year the developer-farmers confronted their deepest fear: legacy systems.

In the tech world, legacy code is old, poorly documented software that nobody wants to touch but everybody depends on. In the farming world, the equivalent turned out to be… well, everything. Soil that had been farmed for generations carried the decisions of every farmer who came before. Irrigation systems built in the 1970s still watered the fields. Barns constructed before anyone alive was born still housed the animals.

“I thought I left legacy code behind,” Sarah Kim wrote in what had become an annual blog post (her writing had become popular enough that she had a book deal, tentatively titled “Microservices and Macro-veggies: One Engineer’s Journey from AWS to Actual Weather Service”). “But it turns out every farm is a legacy system. The original developer is dead. The documentation is oral tradition. The tests are whether your crops survive winter. And the deployment pipeline is a prayer and a tractor.”

* * *

The metaphor was apt, and it led to one of the more surprising developments of 2031: developer-farmers started applying software preservation techniques to farming knowledge.

A group of former documentation engineers—people who, in their previous lives, had written API docs and README files—launched a project called “The Farm Archive.” Its mission was to document the knowledge of traditional farmers before it was lost. They conducted video interviews, created searchable databases of farming techniques, and—in a move that would have been unthinkable five years earlier—published it all as an open-source “FarmBook” that anyone could contribute to.

Earl was one of the first contributors. His entry, titled “How to Tell If It’s Going to Rain (No Sensors Required),” became the most-read page on the site. It contained advice like “If the cows lie down, it’s gonna rain” and “If your knees hurt, it’s gonna rain” and “If I say it’s gonna rain, it’s gonna rain. I’ve been doing this for fifty years and my accuracy rate is better than your fancy algorithms.”

A data scientist actually tested this last claim. Earl’s rain predictions had a 78% accuracy rate, compared to the local weather app’s 82%. Given that Earl’s method required zero electricity, zero internet connection, and zero venture capital funding, this was widely considered a win for Earl.

* * *

On a personal level, 2031 was the year the developer-farmers stopped being developer-farmers and started being just… farmers. The hyphen, both literal and metaphorical, began to dissolve.

Marcus and Earl had become genuine friends—an unlikely pairing of a former Silicon Valley engineer and a septuagenarian Iowa farmer, united by a shared love of growing things and a shared exasperation with the weather. They had a standing Tuesday dinner at Earl’s house where they would argue about everything from soil composition to whether Die Hard was a Christmas movie. (Marcus said yes. Earl said it was “a movie about a man in a building” and refused to engage further.)

Tom’s dairy farm was thriving. The cow named Rust had been bred, and her calf was named Rust 2.0, a name that the actual Rust programming community found either hilarious or insulting depending on who you asked. PHP was still the most productive cow and had somehow outlived every prediction of her demise, which Tom considered the most realistic simulation of the actual PHP language he had ever encountered.

Sarah had dismantled her microservices farm architecture and replaced it with what she called a “monolithic approach”—one big, interconnected garden where the plants supported each other. “It turns out,” she wrote, “that millions of years of evolution already solved the distributed systems problem. It’s called an ecosystem. And the latency is incredible—zero milliseconds between the bee and the flower.”

CHAPTER SEVEN

2032: return 0;

* * *

Six years.

Six years since Marcus Chen typed “I’m done” into a Slack channel at 3:47 in the morning. Six years since the first developer looked at a field and saw not emptiness but possibility. Six years of crashed drones, over-engineered irrigation systems, cows named after programming languages, and goats with an insatiable appetite for both sprint boards and investor briefcases.

By 2032, the Great Migration had fundamentally changed both industries. Tech companies, hemorrhaging talent to agriculture, had been forced to rethink everything—work-life balance, remote work, the very nature of what it meant to build something. Some of the best engineers in the world now spent their mornings in the field and their evenings writing code, and they were happier than they had ever been writing code all day.

Rural communities had been revitalized. Towns that were dying had been reborn. Main streets that were empty now had coffee shops (they couldn’t help it; the developers had to have their coffee shops) alongside feed stores. Schools that were closing had new students. The cultural divide between rural and urban America, while far from healed, had narrowed in ways nobody predicted.

* * *

On a warm evening in October 2032, Marcus Chen sat on his porch overlooking his farm. It was harvest time, and the fields were heavy with produce. The setting sun painted everything in shades of gold and amber that no CSS gradient could match, though Marcus—old habits dying hard—had once tried to find the hex code.

Earl sat beside him, as he did most evenings these days. They were drinking lemonade that Marcus’s daughter Maya had made, using lemons from the small grove that Marcus had planted in his greenhouse three years ago.

“You know,” Marcus said, “when I started this, I thought farming was just… deploying code to dirt.”

“Mmhmm,” Earl said, which was his way of saying approximately forty-seven things simultaneously.

“But it’s not that at all. It’s more like…” Marcus paused, searching for the right words. “It’s more like being in a conversation. With the land. With the weather. With the plants. And the trick is to listen more than you talk.”

Earl took a sip of lemonade. “That,” he said, “is the first smart thing you’ve said in six years.”

Marcus laughed. From somewhere behind the barn, they could hear Maya and Earl’s grandchildren playing, their laughter mixing with the sounds of crickets and the distant hum of a drone returning to its charging station.

“Do you regret it?” Earl asked. “Leaving all that behind?”

Marcus thought about it. He thought about the standing desks and the stand-up meetings. The deployment anxiety and the dependency hell. The endless scroll of Slack messages and the quiet terror of a 3 AM PagerDuty alert. And then he thought about the feel of soil between his fingers, the smell of rain on dry earth, the indescribable satisfaction of watching something you planted push through the ground toward the sun.

“Earl,” he said, “I didn’t leave anything behind. I just finally started building something real.”

Earl nodded. They sat in comfortable silence, watching the sun set over fields that were once just dirt and were now, through some miracle of stubbornness, curiosity, and an absolutely unreasonable number of soil sensors, something like home.

* * *

Somewhere in the house, Marcus’s phone buzzed. A notification from the Devs Who Farm Discord, which now had 3.2 million members. Someone had posted a message that Marcus had seen a thousand times before, in a thousand variations, from a thousand different people standing at the same crossroads he had stood at six years ago:

“I’m done. I’m buying a farm. I’m going to grow things. Real things. Things you can eat. Things that don’t require a YAML file.”

Marcus smiled, set down his phone, and watched the stars come out over his farm.

The circle was complete. The program had finished executing. And the return value, for once, was zero—which, in programming, means everything went exactly as it should.

EPILOGUE

Five Years Later

* * *

In 2037, Earl passed away peacefully in his sleep at the age of eighty-four. His funeral was attended by over two thousand people, including three hundred developer-farmers who credited him with teaching them everything they knew about growing things.

Marcus delivered the eulogy. He spoke about patience, about wisdom, about the way Earl could predict rain by looking at the clouds and predict a bad harvest by looking at the soil and predict a person’s character by watching them work.

“Earl never wrote a line of code in his life,” Marcus said. “He never deployed an application or closed a Jira ticket or survived a sprint review. But he understood systems better than any engineer I’ve ever met. He understood the system of the land, the system of the seasons, and—most importantly—the system of people helping each other figure out how to grow things.”

Marcus paused. “He also had strong opinions about JavaScript, and I want the record to show that the cow we named JavaScript eventually calmed down, which Earl said was proof that ‘even the worst of them can be redeemed.’ I choose to believe he was talking about the cow.”

Laughter, through tears.

Earl’s farm was inherited by his grandchildren, who had grown up alongside Maya and the other children of developer-farmers. They were a new kind of farmer—one that Earl would have approved of, even if he never would have said so.

They kept his methods. They used the new tools. And they named their first tractor “Earl,” because some legacy systems deserve to run forever.

* * *

THE END

// exit code: 0

* * *

ACKNOWLEDGMENTS

To the developers who dream of dirt.

To the farmers who tolerate them.

And to the cows who, against all odds,

respond to the name PHP.