Graph the True Data and Reality

I finally read Data and Reality by William Kent. A classic. A necessary read. A slow one, intentionally so.

I saved the last 80 pages for a train ride from the Midlands to London Euston, and I’m glad I did. There was something poetic about moving through space while reading a book that questions how we represent movement, relationships, and meaning in structured data.

I always suspected this book would change how I see things. And it did.

You see, I’ve been dancing in and out of the buzz, like most of us in tech. Sometimes I wonder if I’m heading in the right direction, especially when everyone’s shouting about the next best framework, architecture, or LLM breakthrough. But this book slowed me down in the best way. It reminded me that underneath all the noise, there’s still a deep and unsolved question:

How do we represent reality with data, without flattening it?

Kent doesn’t offer perfect answers. But maybe that’s what makes the book so good. It asks better questions than most people are even thinking about.

And for me, it’s brought my attention back, again, to graphs.

Because, if you think about it, graphs might be closer to the real world than any other model we have. Not perfect, no. But closer.

Here’s where it started to make more sense.

Kent points out that most data models, especially the logical ones, aren’t designed to reflect human reality. They’re designed to reflect data processing requirements. Things like entity-attribute-value triples, naming conventions, and record locations. Useful, yes. But inherently biased toward how computers store and access information, not how people understand or relate to the world.

“Most models describe data-processing activities, not human enterprises.”

, William Kent, Data and Reality

That line stuck.

Because when you think about a person, you don’t think of them as rows in a table. You think of them in motion. With evolving relationships. Changing addresses. New titles. Old titles. Linked memories. Conditional truths.

Take something as simple as an address. A person lives somewhere. Then they move.

In a relational model, you’re probably updating a field. In reality? That address isn’t invalid, it’s just a past relation. Maybe relevant. Maybe not. But it existed.

In a graph? You don’t overwrite, you link. You preserve the relationship with time, context, and perhaps even purpose. You don’t erase, you connect.

And isn’t that how life works?

The more I think about it, the more I feel graphs, especially semantic graphs and knowledge graphs, have the potential to mirror real-world complexity in a more forgiving, flexible, and honest way.

But then again, Kent also reminds us that no model is perfect.

Confusion when modeling the real world? That’s a sign you’re paying attention.

Just don’t get stuck in it. Time-box it. Let the ambiguity exist. Then move forward.

“All models are wrong, but some are useful.”

, George Box (not from the book, but a helpful frame)

The key, according to Kent, is context.

What is the purpose of your model? That question alone should shape how you represent your entities and relationships. The more grounded your purpose, the more useful your abstraction.

So… back to the question I’m asking myself:

Is graph the answer?

Maybe.

But maybe the question isn’t “Is graph the closest reality to human information retrieval ?”

Maybe it’s: There isn’t any perfect solution, just something close. something that can help us capture ambiquities and improve discovery of our reality*?*

Maybe that’s where graph thinking becomes compelling, not because it solves all ambiguity, but because it leans into it. It accepts that people, places, events, and meaning are rarely fixed in neat boxes. We change. We relate. We move. A person moves from one city to another, does their “address” change, or is it just another chapter in their story? Do we overwrite the old or relate it to the new?

These aren’t just data questions, they’re philosophical ones. And that’s the beauty (and challenge) of modeling reality. You’re not just structuring data. You’re interpreting life.

So maybe the question isn’t “Is graph the solution?” but “Can anything else hold this much nuance?”

And if it can, how do we shape it so it still makes sense to a machine?

Because in the end, modeling reality isn’t about getting it perfectly right. It’s about getting close enough to act meaningfully. To see just enough of the pattern to take the next step.

That’s the thing with graphs. Once you start thinking in entities, relationships, and attributes, you start seeing them everywhere. In projects. In ideas. In yourself.

So yeah, Data and Reality didn’t give me a roadmap. But it gave me something better: a mirror. A way to reflect on how I think about data, structure, and meaning.

And maybe that’s what I needed, to go deeper into graph theory, linked data, semantic models, contextual graphs, and inference engines.

Because lately, I’ve been thinking about where I want to spend the next 5–10 years of my work. What kind of problems I want to live inside of. And right now, it’s this one:

How do we represent reality, and deal with its nuance, in a way that both humans and machines can understand?

That means graphs. That means semantics. That means knowledge graphs. It probably means language models and vector data stores too, especially as we explore how to retrieve meaning from information, not just store it.

But even then, graphs and vectors, reasoning and embeddings, it all circles back to the same thing: Trying to make sense of messy, beautiful, contradictory reality.

I don’t have answers. I just know this book made me sit with the questions longer. And maybe that’s what I needed.

To keep asking questions like:

• How do we preserve truth over time?
• How do we model uncertainty without collapsing it?
• And how do we make meaning portable across systems and people?

It’s still a lot.

But it feels like the right kind of a lot.

And maybe this is what getting closer to the work really looks like. Because, life and reality in itself can be a lot.

,

Let me know what you think.

Or maybe… this is just another day.