The first architectures of the Judicial Server took shape in 1991. The legal and commercial chronology of the system would then unfold through several stages.

Until then, we had learned how to circulate information, structure it, organize thesauri, and build architectures capable of handling major surges in demand. With the Judicial Server, we were entering a far more sensitive territory: one where information directly encounters justice, struggling companies, public sales, and price formation.

The surviving documentation confirms this chronology and shows that, by the early 1990s, the Judicial Server already had structured judicial, economic, and regulatory databases, with nomenclatures, indexing, and telematic distribution.

Yet the problem that interested me was almost brutally simple.

An auction could be legally public while remaining, in practice, economically reserved for a handful of insiders.

I found that contradiction intolerable.

When a company entered judicial liquidation, its assets had to be inventoried. For smaller businesses, the inventory could remain relatively limited. For an industrial company, it took on an entirely different scale.

Hundreds of pages. Sometimes thousands. Machine tools, computer equipment, technical installations, furniture, inventory, specialized equipment, serial numbers, specifications, accessories, technical documentation: before the company disappeared as an economic entity, the inventory report produced what was almost its final anatomy.

Our documentary corpus describes this function precisely: the report could become a genuine industrial, economic, and sociological X-ray of the company in liquidation.

I had a friend who was a judicial auctioneer and who could spend fifteen days in a large factory carrying out this work.

Everything had to be seen.

Everything identified.

Everything itemized.

There was nothing bureaucratic about this degree of precision. It was necessary so that the assets could subsequently be sold.

And yet, when notice of the sale finally reached the public, what sometimes remained of this immense body of documentary intelligence?

A few lines.

“Extensive computer equipment.”

“Office furniture.”

“Industrial equipment.”

“Robotics.”

“Various equipment.”

In other words, the complete document existed. But the potential buyer did not see it. This is where I encountered what I would spend my entire life calling information asymmetry.

Some people knew. The others knew only that a sale was going to take place. Between those two levels of knowledge, an enormous difference in value could emerge.

A buyer who knows the exact model of a machine, its condition, its location, and its equipment may decide to travel five hundred kilometers. Someone who merely reads industrial equipment stays home. He does not lack motivation. He lacks information. And in this world there existed what was known as the bande noire, the “black band.”

The term referred to a small circle of professionals who, because of their proximity to the sales, the clerks, or the various actors in the system, were able to obtain infinitely more precise knowledge of what was going to be sold. They did not arrive at the auction discovering the lots. They knew. They could prepare. Compare. Evaluate. Position themselves.

I do not want to turn that nickname into a general accusation against an entire profession. That would be false and unfair. It describes an information system in which a handful of insiders could enjoy a considerable advantage over the ordinary public.

That was the system I wanted to break. Not the auctioneers. Not judicial sales. The information privilege.

I therefore never regarded regulated professions as adversaries in themselves. Our documents correct an earlier counting error here: I did not meet around one hundred auctioneers, but approximately 550, after several years of fieldwork and hundreds of thousands of kilometers traveled.

Three hundred and sixty thousand kilometers.

You have to understand what that distance represents. I did not stay in Lyon behind a screen explaining to auctioneers how their profession worked.

I went to see them.

In their offices.

At the sales.

In the factories.

In midsize towns.

In the major centers.

I wanted to understand their sociology, their law, their vocabulary, their economic constraints, their relationship with the public, and the reasons why information circulated the way it did.

That has always been my method. Before claiming to change a system, you have to pass through it.

That immersion taught me something important: auctioneers had constraints of their own. Public notice costs money. Advertisements have a price. A judicial sale involves deadlines and obligations. It would have been intellectually dishonest to attribute every shortcoming to a deliberate desire for opacity.

But acknowledging those constraints did not change the underlying problem.

If the sale was public, the information enabling the public to participate in it had to be able to circulate far more widely. And this is where the law comes in.

I knew that reports, inventories, and various forms of judicial information could not simply be treated as private property to be disposed of at someone’s discretion.

The corpus of our interviews summarizes what followed: legally public information, inadequate material publicity, asymmetry between insiders and the public, recourse to the law to obtain disclosure, then transformation of the static document into a searchable database.

At one point, the conflict became direct. I remember an auctioneer making it very clear to me that we had no business interfering in matters that did not concern us.

My answer was equally clear. If this information concerns a public sale and if the ministerial officer holds it in the course of his official duties, then whether it can be disclosed is precisely our concern.

I was not circumventing the system. I was asking it to apply its own rule.

The tone rises.

The law is invoked.

Formal reports are prepared.

And then comes a scene I have recounted several times because it condenses almost that entire period.

The fax machines start spitting out pages.

One page. Then ten. Then a hundred. Then masses of documents. The inventories arrive.

What only hours earlier we had been told did not have to be disclosed to us suddenly begins pouring into our machines.

I watched the faxes come out and understood that we had crossed a boundary. But obtaining the information was still only half the work. Receiving a thousand pages does not mean producing a thousand pages of knowledge. You can very easily replace legal opacity with documentary chaos.

We had to enter the data. Break it down. Qualify it. Index it. Link it. Build searches.

A machine had to be distinguished from a vehicle, from computer equipment, from furniture, from inventory.

The sale had to be associated with a date. A jurisdiction. An auctioneer. A company. A location. A proceeding.

The information had to become navigable. This is where the full power of the ALJ software — Annonces Légales et Judiciaires — appears.

The document preserved in the library materially shows this telematic architecture: ALJ was developed to process legal and judicial notices through an organization based on categories, consultations, subscriptions, and specialized groupings.

ALJ should not be understood as a simple data-entry program. It is one of the earliest manifestations of our doctrine of verticalization. From masses of judicial, economic, and legal information, we built specialized databases. We had approximately 72 verticalized databases within this architecture.

That two-page spread Nadège placed back in front of me decades later is extraordinary. It shows how we thought. The last thing we wanted was one gigantic bin filled with notices. We wanted to understand the nature of each piece of information.

The creation of a company is not a liquidation. A change in corporate bylaws is not a ban on managing a company. A judicial sale is not a voluntary sale. A real-estate notice is not a public procurement contract. Data exists properly within a system only when it knows what it is.

That sentence sounds almost philosophical. It is simply industrial. If you mix information governed by different regimes, you lose its meaning.

ALJ was therefore beginning to produce something that, to my knowledge, no one had yet structured in this way and on this scale.

A national geography of legal and economic events.

And this is where the roughly five hundred to five hundred and fifty legal-notice newspapers take on their full significance. France was covered by authorized newspapers, sometimes several within the same département. Each had its own publication schedule. Its own layout. Its own formulations. Its own rhythm. Its own territory.

Taken separately, they were perfectly useful. Taken together, they formed a fragmented national memory.

Who could read five hundred newspapers every day and reconstruct the state of the French economy?

No one.

A man could be banned from managing a company in Nice and have interests in, or have managed companies in, Lyon or Paris. The information was public. But it was localized.

Anyone who did not know that he needed to look in Nice would never find it. Here was another information monopoly with no explicit owner. It arose simply from fragmentation.

So we began to aggregate. The newspapers. The court registries. The publications. The notices. Then we made them searchable within a national architecture. The local was no longer trapped within the local.

Information could now be linked to a person, a company, or another event located hundreds of kilometers away. That is an enormous change.

The database did not merely make the information easier to read. It created knowledge that had not existed when the documents remained separate.

The relationship produced new information.

That discovery directly prepared the way for Artprice. Above all, the Judicial Server taught me that price also depends on the circulation of information.

Price is never merely the result of a hammer falling. Price depends on the number of people who know that the sale exists. On the number of people who know enough about the lot to decide to come. On the level of confidence they have in the information. And therefore on the quality of the publicity.

Before our intervention, some industrial sales could bring together an extraordinarily small number of genuinely informed buyers.

Eight.

Nine.

A few regulars.

And in front of them, sometimes an entire factory.

Thousands of lots.

Tens of thousands of individual items.

At the time, I considered that the asymmetry could reach not merely a ratio of one to ten but, depending on the circumstances, one to one hundred or even one to one thousand. I give these orders of magnitude as my assessment at the time. But the logic is simple.

One man had one hundred pages of information. The other had four lines. They were not operating in the same market.

When we began distributing much more precise information, the number of potential buyers increased. Bidding became more competitive. Prices could rise. And I saw situations in which better publicity very substantially transformed the proceeds from the sale of the assets.

That matters legally and economically. Creditors recover more. The proceeding recovers more. But there was also something human that financial tables do not show.

Business owners would tell me:

“Thank you. You saved my honor.”

I understood that sentence.

A company in liquidation is not merely a case number. For the person who built it, it may represent twenty or thirty years of life.

Employees. A family. Nights. Loans. Expertise. Machines purchased one by one. Bankruptcy is already an act of violence. Then comes the sale.

When the assets are dispersed among a handful of perfectly informed buyers and sell at prices the owner considers derisory, he may experience that scene as a second humiliation.

I used to say that some companies were struck twice. The first time by bankruptcy. The second by undervaluation caused by inadequate circulation of information.

This idea is preserved very clearly in our interviews: the problem was not the sale itself, but the economic justice created or destroyed by asymmetry of information. When more buyers arrive and the price rises, the company obviously does not come back to life.

But something is restored. The value of what had been built is better recognized. Creditors recover more. And sometimes a man walks away with the sense that his work was not entirely liquidated like a pile of scrap metal.

That was when computing became, for me, a question of justice. Not justice in the abstract sense. Concrete equality before information.

I believe that this is also where my conviction arose that transparency does not necessarily destroy a market. On the contrary, it can expand it. We observed this.

More information produces greater trust. Greater trust can attract more participants. More participants produce greater competition. And competition gives price a different quality.

I later formulated this intuition in an almost provocative way: “When an opaque market becomes symmetrical, it does not collapse; it can explode.”

Our working archives retained this sentence because it directly connects the Judicial Server to Artprice.

The Art Market would confront me with exactly the same question a few years later. A handful of insiders know the prices. A handful of experts possess the historical records. A handful of professionals know that an artwork has already appeared at auction three times. Everyone else buys with partial information.

What happens if that memory becomes accessible?

Some thought the market would collapse.

I had already learned in the judicial world that the answer could be exactly the opposite. Transparency can produce liquidity. Because it produces trust.

That is why the Judicial Server is much more than a technical ancestor of Artprice. It is its philosophical laboratory.

It teaches me that publicity is a democratic function. Not publicity in the commercial sense. Publicity in the legal sense: making visible what must be capable of being invoked and accessed. Publicity of legal acts. Publicity of sales. Publicity of knowledge.

I then understood that economic freedom itself depends in part on the quality of that publicity. You cannot choose what you do not know exists. You cannot bid on a machine no one has described to you. You cannot properly assess an asset when some people know its history and others know only its name.

This philosophy of publicity would become a permanent thread. It explains why, much later, I would so often refuse requests to erase an auction result from Artprice simply because it had become inconvenient.

A public event has produced a trace. The memory of that trace belongs to the history of the market. The principle is already there, in judicial sales.

Public information does not become private again because it inconveniences someone. But the Judicial Server did not merely change sales.

It also confronted us with the infrastructures themselves. When publicity improves, more people consult the system. When more people consult it, the machines suffer.

We experienced those extremely abrupt surges in demand when engineers would naturally have proposed adding still more computing power. I began thinking differently.

The problem was not always absolute power. It could be the brutality of the incoming load.

It was in this environment that we worked on what I called fantomisation: absorbing, distributing, and delaying certain loads so that a massive influx would not strike the entire architecture simultaneously.

Here again, I recognize my way of thinking. When matter resists, I do not necessarily try to oppose it with greater force. Sometimes I try to change the geometry of the problem.

This experience would prove decisive for the architectures of Groupe Serveur and later Artprice.

Hundreds of thousands of potential users could arrive after a television campaign or a mass distribution. The system had to hold.

Once again, commercial success interested me only secondarily. What mattered was the learning. We now knew how to build a database.

Feed it. Make it national. Secure it legally. Open it massively to the public.

Then absorb the technical consequences of that opening.

The Judicial Server also gave us new legitimacy among regulated professions and major economic actors.

Our working documents had rightly reformulated the point this way: rather than speaking of prestige, it is more accurate to say that this experience definitively established Groupe Serveur’s credibility in fields where data was legally complex, regulated, and defended by powerful interests.

That credibility would enable us to scale up. Because the next problem was already visible. France was no longer enough. Law circulated. Companies circulated. Sales circulated. Europe’s economic borders were becoming increasingly porous.

A French executive could operate in Belgium. A German company could buy in France. Judicial information produced in one state could become relevant in another.

Our national architecture was therefore inevitably going to reach its limit. And I already knew what would happen. After learning to cross départements, we would have to cross states. After normalizing French information, we would have to learn how to make different laws, languages, nomenclatures, and administrative cultures communicate with one another.

The Judicial Server had taught us how to break an information monopoly.

The European Judicial Server would force us to do something far more difficult: scale up without losing the meaning of the data.