The experiment (the “Pottery Class Paradox”)

The story goes like this: a ceramics teacher splits a class into two groups.

• The first group was graded only on quantity. Their final grade depended on the total weight of pots they produced.
• The second group was graded on quality. They had to create one perfect pot.

At the end of the semester, the best pots came from the quantity group. Why?

• Not because they cared less.
• Because they practiced more.
• They made mistakes.
• They adjusted their technique.
• They learned from each failed attempt.

The quality group spent most of its time thinking about perfection.

Why does this matter for architecture?

This story is a perfect metaphor for evolutionary architecture.

In software, the quality group mindset looks like this:

• Big Design Up Front
• Long architecture documents
• One ideal target state
• A rewrite plan that must be right

The quantity group mindset looks different:

• Ship small increments
• Observe real behavior
• Refactor safely
• Improve continuously

The best architectures rarely emerge from a single brilliant design session.
They emerge from controlled change.

Evolutionary Architecture in simple terms

Neal Ford describes evolutionary architecture as designing systems that support incremental and guided change as a first principle.

That phrase sounds complex. The idea is not.

Build systems that can evolve.
Change them often.
Measure if they stay healthy. But how should it be measured?

Fitness functions are the kiln

In the pottery story, the kiln reveals the truth. A pot either survives heat or it cracks.

In software, production is the kiln.

This is where fitness functions come in. They are automated checks that protect key architectural qualities:

• performance
• security
• reliability
• cost
• resilience

They act like a scale in the pottery class. They tell you if your system is getting better or worse.

You do not hope the architecture is good.
You measure it continuously.

Architecture is not a statue

A statue is finished. It cannot adapt.

A good architecture is closer to a living system. It grows. It changes shape. It responds to stress.

If you try to design the perfect system on day one, you are acting like the quality group. You are optimizing for elegance instead of learning.

Evolutionary architecture optimizes for learning under constraints.

It means:

• short feedback loops
• safe experiments
• strong guardrails
• continuous refinement

Quality is not the starting point.
It is the result of disciplined iteration.

Like pottery, great architecture is shaped under heat.