Blog #1 — 11 October 2026

I’ve spent a considerable amount of time thinking about how systems work.

In engineering, the starting point is usually a specification. We agree what something is supposed to do, define how its components interact, and establish rules that make the outcome predictable.

A USB device works because manufacturers agree on a standard. Software communicates through protocols. Machines can be tested against known requirements.

When something goes wrong, we can ask a relatively straightforward question: did the system behave according to its specification?

But lately I’ve been wondering what happens when we apply that same way of thinking to human beings.

Because the law is also a system.

It has rules, procedures, classifications, deadlines and mechanisms for enforcement. In theory, these allow millions of people to live together under a shared framework.

The difference is that human lives don’t arrive with specifications.

A World That Doesn’t Follow the Rules

An engineer can design a system around known inputs.

A legal system has to deal with memories, intentions, misunderstandings, changing circumstances, missing information and people who don’t necessarily behave as expected.

Two people might witness the same event and remember it differently. Someone might move house and never receive a letter. A vehicle might be registered to one person, driven by another and left in the possession of a third.

The administrative system still needs to produce an answer.

It needs names, dates, addresses, responsibilities and outcomes.

But the fact that a situation can be reduced to fields in a database doesn’t mean those fields tell the whole story.

A record can be complete according to its specification and incomplete according to reality.

That’s where the engineering analogy begins to get interesting.

A Mustang on Judge Judy

I was reminded of this while watching an old episode of Judge Judy involving a Ford Mustang and a test drive that ended badly.

What interested me wasn’t simply who damaged the car. It was the way the case exposed different layers of responsibility.

There was the person who drove the vehicle, the person who owned it, the costs of recovering it and the question of what the car was actually worth.

From a purely mechanical perspective, the problem seemed simple: a car had been damaged.

But legally, the answer depended on several different questions. Who had the relevant rights? Who was responsible for particular decisions? What loss could actually be established?

At one point, the discussion turned to the car’s surprisingly low value and the towing and impound costs.

It struck me that the physical event and the legal consequences were not necessarily the same thing.

A damaged car is an engineering problem.

Determining who should pay, and how much, is a problem involving evidence, responsibility and judgment.

The law has to turn a complicated human event into a decision that can be enforced.

And that transformation isn’t always straightforward.

The Difference Between a Rule and Reality

In engineering, rules are often agreed before the system is built.

If two components don’t implement the same protocol, they may not communicate. If a machine receives an invalid input, we can design it to reject that input.

Human society is different.

The law applies whether or not someone has read every relevant statute or understands every procedure.

There are good reasons for this. A society couldn’t function if each person could choose which laws applied to them.

But it creates an unusual relationship between the system and its users.

We are expected to comply with rules that may be complicated, distributed across multiple institutions and difficult to understand without specialist knowledge.

And unlike a piece of software, we cannot simply inspect the entire specification before participating.

We’re already participating.

When Institutions Become Components

Modern legal administration increasingly resembles a network of interconnected systems.

A court makes a decision. Another organisation records its consequences. An enforcement process may then act on that record.

Each institution has its own responsibilities, procedures and data.

From an engineering perspective, this sounds familiar. It’s a distributed system.

And distributed systems introduce a particular problem: one component may trust information received from another without independently verifying how that information was established.

Imagine an error entering at the beginning.

The first institution records it. The second receives it. The third acts on it.

By the end, several organisations may possess records that appear to support the same conclusion.

But those records aren’t necessarily independent evidence.

They may all be copies of the same original assumption.

A mistake doesn’t become true because it has been replicated successfully.

That seems obvious when talking about software.

It becomes considerably more important when the records concern human beings.

The Problem With Perfect Compliance

There’s a temptation to judge a system by whether it follows its own rules.

Did the correct form get issued? Was the record transmitted? Did the next department apply the prescribed procedure?

Those are necessary questions.

But they aren’t sufficient.

An institution might follow its procedures correctly while acting on incomplete or inaccurate information.

This doesn’t mean procedures are pointless. Quite the opposite: procedures protect people from arbitrary decisions.

It means the quality of a system depends not only on its normal operation but also on what happens when its assumptions are challenged.

Engineers understand this distinction.

We test for failures. We build diagnostics. We preserve logs. We design recovery mechanisms.

We don’t simply declare a machine reliable because it performs well when every input is correct.

We ask what happens when something unexpected occurs.

The legal system deserves the same scrutiny.

Can We Engineer Better Justice?

My first instinct when I look at administrative complexity is to ask why more of it can’t be automated.

Why should obtaining a record take weeks? Why should the same information need to be requested from several organisations? Why should someone have to navigate telephone menus just to find out which institution holds a particular document?

Surely technology could make these things easier.

And perhaps it can.

But I’ve also begun to wonder whether automation introduces a different danger.

If a decision is wrong, making its consequences faster and more efficient isn’t necessarily an improvement.

A system that can automatically impose a consequence should ideally be equally capable of explaining the decision, preserving its supporting evidence and providing an effective route to challenge it.

Otherwise, we risk building systems that are exceptionally good at enforcing their conclusions and surprisingly poor at reconsidering them.

That is not an argument against digitisation.

It’s an argument for thinking more carefully about what we’re digitising.

Why I’m Writing This

These questions are no longer entirely theoretical for me.

I’m preparing to navigate a legal and administrative process concerning records and decisions that affect me personally.

Over the coming days, I intend to document that journey here.

I’ll write about the records I request, the responses I receive, the procedures I encounter and the practical lessons I learn.

Some of my assumptions may turn out to be wrong. Some apparent inconsistencies may have straightforward explanations. Others may prove more difficult to resolve.

I’ll try to distinguish carefully between documented facts, personal recollections and unanswered questions.

This isn’t legal advice, and I don’t intend to turn an unresolved situation into a collection of accusations.

My hope is that a transparent account of the process might help other people facing similar circumstances.

And perhaps, along the way, it will reveal something useful about how our institutions work.

The Specification Is Not the Whole Story

Engineering has taught us extraordinary things about reliability, consistency and the value of shared rules.

But human beings are not machines.

Our circumstances are messy. Our memories are imperfect. Our intentions aren’t always visible. And the information recorded about us isn’t always the whole truth.

A legal system needs rules to function.

Yet if those rules are to serve justice, the system must also have room for the unexpected, the disputed and the genuinely mistaken.

Because a system isn’t truly reliable simply because it follows its specification.

It must also be capable of recognising when the specification fails to describe reality.

And, more importantly, it must know what to do next.