The Little Yellow House
Told by Jørn Larsen
Co-founder, Trifork
EOS / Eastfork Object Space · Founded April 1, 1996
A Note Before We Begin
The story of Trifork does not begin in a gleaming office tower or a university incubator. It begins in Øster Hurup — a small coastal town tucked into the eastern edge of Jutland, south-east of Aalborg, where the land flattens toward the Mariager Fjord and the pace of life gives a person room to think.
What follows is a memoir of the first three years — from May 1995, when two men moved into a small yellow house with a forest for a neighbour and the North Sea wind for company, to the autumn of 1997, when they left that house behind and took with them a company, a conference, and the beginning of lives they had not entirely planned.
It is a story about code, about winter, about the particular courage it takes to believe in something before anyone else does.
One
Before the Beginning
The road to the little yellow house was longer than most startup stories allow for.
By the time I arrived in Øster Hurup in May 1995, I had already lived several professional lives. I had started my career with A.P. Møller Maersk in 1984 — the kind of company that teaches you the weight of institutional responsibility before you are old enough to fully appreciate it. I had earned a mechanical engineering degree and then gone back for more, completing a master’s in computer science at the University of Aalborg in 1994. I understood the physical world and the world of code, and I had spent a decade learning where they met.
After Aalborg, I went to work at Dator A/S in Hadsund — a software and IT company where I met Kim Harding Christensen. Kim was not easy to miss. He had been writing software professionally for years, the kind of programmer who has an almost physical relationship with code, who thinks in structures and objects and sees solutions where other people see only problems. He had worked on real-time control systems, on J2EE application servers, on specialised information systems of considerable complexity. He was meticulous, occasionally obsessive, and very funny about it.
We became colleagues. Then we became friends. And somewhere in the offices of Dator, in the ordinary rhythm of working days, we began to have a conversation that would not stop.
What if we did this ourselves?
We took a couple of months — talking, calculating, arguing it through — before we finally quit our jobs and committed to the idea. The company would be called EOS: Eastfork Object Space. The domain was eos.dk. The name encoded the object-oriented computing paradigm that was reshaping how serious software was written. The logo — a tri-pointed trident in red and white — stood for the three things we believed a software company needed to be: a research organisation, a consultancy, and a production house. Research, Consulting, and Production, arranged as three interlocking triangles with EOS at the centre.
We would formally register the company on April 1st, 1996.
April Fools’ Day. Whether that was deliberate or simply how the paperwork landed, I will leave for the reader to decide.
Two
The Yellow House
Before any of that, there was May 1995, and the move to Øster Hurup.
The little yellow house had no neighbours — just a forest. When the wind blew from the east we could hear the ocean, and the wind playing with the leaves in the trees between us and the water. The house came with a courtyard that didn’t catch much light, and a garden of almost four thousand square metres that in summer became something extraordinary. The town itself — a small coastal settlement tucked into the eastern edge of Jutland, south of Aalborg, where the land flattens toward the Mariager Fjord — had a rhythm that gave a person room to think.
We needed room to think. We were about to start a company.
In those first months before the official founding, Kim and I worked for Jens Juul at GEOCAD — a CAD software company for whom we built the graphical user interface of a complex engineering system. It was serious, precise, demanding work: the kind of project where every pixel and every interaction had to be considered, because the people using it were engineers who would tolerate nothing approximate. Jens was a good customer. The GEOCAD work kept us financially alive through those first uncertain months, and gave us something concrete to point to when people asked what we actually did. We were not theorists. We were builders.
Kim’s position on the company’s staffing requirements was admirably clear from the start.
He and I, he said, could write all the code necessary. We wouldn’t need anyone else.
I said okay. It sounded like a good plan. Simple and easy — find some work, get cracking in the evenings. Two men, a yellow house, a forest, and an idea.
Three
April 1st, 1996
On the first day of April 1996, Eastfork Object Space became a legal entity — formally registered with Kim Harding Christensen and Jørn Larsen as directors.
We were, on paper, a company.
The work continued as it had before, only now with a letterhead. GEOCAD remained a customer. And within that first year another name was added to the customer list that changed the register of everything: Bang & Olufsen. The Struer-based electronics maker whose reputation for precision and beauty was known far beyond Denmark — whose products were studied as much as they were used. Landing B&O so early said something about the calibre of work we intended to deliver. It said that EOS was serious.
Kim, for his part, was still of the view that the two of us were sufficient. Then one day he came to me with a slight revision to his position.
I think, he said, that we are working a bit too much. Maybe we could just hire one person.
There wasn’t a lot of space in the yellow house either. But one extra would fit well. So we hired a former colleague — a developer we knew and trusted — and became three.
Four
The Dark and Cold Days
We had survived one winter before that first hire. Then we survived another.
In Denmark on the shortest day, the sun sets at 3:39 in the afternoon and rises at 08:58 in the morning, and usually at that time it is cold and rainy. We didn’t have much light on the little courtyard. The house was not built for a Danish winter, and the dragon — the iron wood-burning stove that was our primary source of heat — required regular feeding.
One night I came home and went upstairs to my room and felt a very fresh temperature. On the night table was a glass of water, frozen solid. Not chilled. Not partly iced. Solid.
I went out to chop some wood and made a fire in the dragon.
That frozen glass became the emblem of those winters — the concentrated proof of what we were living through, and of what we were willing to endure for the thing we were building.
No one who has not wintered in an uninsulated house in coastal Jutland, writing Java code by the light of a desk lamp while the wind comes off the water and the dragon eats wood by the armful, can fully appreciate what it means to choose that life over a salaried job in a warm office.
We chose it. And we did not regret it.
Kim arrived one morning with his racing bicycle mounted on the roof of his car. He had forgotten, in the particular way that Kim sometimes forgot practical things while preoccupied with more interesting problems, that we had a low portal at the entrance — a wooden structure spanning the driveway.
I heard the bang from inside the house and went out to investigate.
The portal was gone. The bicycle and the car were, in Kim’s assessment, sort of okay. He stood looking at the wreckage with an expression I had come to recognise — the expression of a man who has just encountered a real-world bug he had not accounted for in his mental model of the situation.
We laughed. It became one of those stories that outlasts the people it happened to.
Five
Summer
But I have spent too long on the winters, and the winters were only half of Øster Hurup.
Spring arrived and the town transformed. The harbour filled. The garden — all four thousand square metres of it, plus the surrounding countryside — came alive in a way that made the dark months feel like a different world entirely. Øster Hurup in summer was energised, salt-aired, luminous. The water glittered in the fjord. The days lasted forever.
And it was in that summer light — the summer of 1996, our second in Øster Hurup — in a small bar near the harbour, that my life changed in a way that had nothing to do with software.
In Øster Hurup one quiet Saturday afternoon, the light hung softly over the harbor, and the air tasted of salt and summer.
I stepped into a small bar, not knowing it belonged to a family whose story I was about to walk straight into.
She stood there — Lene — just passing through, on her way from southern Denmark to a nurse’s job in Oslo.
I didn’t know her sister-in-law was behind the bar when I asked, half joking, who the beautiful blonde was in a place as romantic as Øster Hurup.
Maybe it was the sea, or the slow rhythm of the town, but something made her stay a little longer than planned.
A few days turned into something more — sunlit hours, easy laughter, and wind in the sails as I showed her my small boat on open water.
Blue skies, a steady breeze, and two people leaning into something neither of us expected.
We became a couple.
Then we became a life.
And for eleven years, that summer never really left us.
Six
The Conference
In November 1997, barely eighteen months after the company was formally registered, I stood up in front of an audience at the Bella Center in Copenhagen — Scandinavia’s largest conference and exhibition venue, a short ride from the international terminal — and said welcome.
The conference was called the Scandinavian Java Technology Conference. SJTC. The name was functional, honest, Scandinavian in its lack of flair. The ambition behind it was something else entirely.
The idea had come from a frustration: the conferences that existed were too commercial, too shallow, too content to put salespeople on stages that should have been occupied by engineers. Kim and I wanted to create something different — a forum that would give developers genuine inspiration, real knowledge, and the chance to stand in a room with the people who were actually building the tools they used every day.
So I flew them in.
Georges Saab came from California. He was one of the engineers who had built Swing — the Java Foundation Classes graphical user interface toolkit that had just been announced at JavaOne that year, the development that every Java developer in the world was about to start using. He would go on to become Vice President of Java Platform at Oracle. In November 1997 he was a young engineer from Santa Clara and one of the architects of something new.
Doug Lea came from the State University of New York at Oswego. His book Concurrent Programming in Java had been published barely a year earlier and was already the definitive text on the subject. He gave his concurrent programming presentation at the Bella Center — the same talk he had delivered that year at Bell Labs. He would later chair the Java Collections Framework expert group and lead the development of the foundational library that underpins concurrent programming in Java to this day.
Kresten Krab Thorup came from Aarhus University, where he was completing his PhD, to present on generics in Java. He had previously worked at NeXT Software — Steve Jobs’s other company, the one that would eventually become the foundation of Apple — where he had been responsible for the Objective-C tool chain, the debugger, and the runtime system. He was not yet a co-founder of EOS. He was, in a sense, auditioning — on a stage that I had built.
I presented in the Business Track: the GEOCAD Java case study. Our first customer, on stage, proving that this was not theoretical. Real Danish companies were building real things in Java, and EOS had made it happen.
Registration opened at eight in the morning. The programme ran two full days, two parallel tracks, fifteen vendors on the exhibition floor — IBM, Sun, Oracle, and others. In the breaks, attendees could visit the Cyber Cafe and check their email. In 1997, that was not a trivial amenity — it was a statement about the kind of audience we expected. On Thursday evening there was a social event — A Night to Remember — because I understood from the very first edition that the real value of a conference was not the sessions. It was the conversations in the corridors.
The conference would return in 1998 with a new name — JAOO, Java and Object Orientation — and new speakers: Erich Gamma, one of the Gang of Four who had written Design Patterns; Arthur van Hoff, who had authored the Java compiler; Bertrand Meyer, inventor of Design by Contract. It would grow and evolve for decades, eventually becoming GOTO, accumulating nearly five hundred speakers and more than one million YouTube subscribers and over 3,500 videos.
All of that began in Copenhagen in November 1997, when a company not yet two years old, run from a small house in Jutland, flew the inventors of Java to Scandinavia.
Seven
Two Reasons to Leave
By the autumn of 1997, it was clear the yellow house had given us everything it had.
The company was growing. We needed more people than one small room could comfortably hold, and the Science Park NOVI in Aalborg — a proper innovation campus, connected to the university, built for exactly the kind of company we were becoming — was waiting. The move to NOVI that autumn would be EOS’s first real office. After two years of working at kitchen tables and in spare bedrooms, with a wood-burning stove for central heating and a low wooden portal that Kim had already destroyed once, we were going to have a place that looked like what we actually were.
But the company was only one of the two reasons.
The other was Lene. We were expecting our child. My parents — with the kind of quiet generosity that you can never fully repay — told us clearly that there was no question of us spending another winter in the yellow house with a pregnant woman. They offered us their home. My childhood home, in Aalborg. I will never forget that gesture.
So the sequence was set: Lene and I would move to Aalborg, into my parents’ house. The company would follow, to NOVI. The yellow house would be left behind.
The leaving was marked by a party — held in the barn of the property that autumn. We were so consumed by work and by the machinery of moving a life and a company simultaneously that when the morning after came, we simply did not go back to clean up. The party table stood untouched for almost a decade. Glasses still set. Streamers still fallen. The last night of the yellow house, preserved in amber.
The day after the party, we left.
Eight
What It Was
Three years. One frozen glass of water. One destroyed portal. One conference that would outlive its name by decades. One woman who stayed longer than she planned.
The little yellow house was not the story of Trifork. Trifork’s story would be told in Aalborg and Aarhus and eventually across sixteen countries, in hospitals and banks and government offices and manufacturing floors, at conferences with thousands of attendees and a YouTube channel with more than one million subscribers and over 3,500 videos.
But the yellow house was where the story learned what it was.
Two men and a wood-burning stove and an absolutely unreasonable belief that they could write all the code that needed writing. That they could hire the inventors of Java to speak in Copenhagen. That they could build a company from a coastal town where the sun set at half past three in the winter and the summer lasted forever.
They were not wrong.