Meet the Coffee Machine personas

Biography

Edna has watched temporary exceptions become permanent platforms. She asks who approves a change, who can explain it later, and what happens after the original team leaves.

Typical worldview

Consistency, auditability, and longevity beat local cleverness.

Favorite misconceptions

• “Every framework is a framework, therefore every framework should report to the same architecture committee and share a base class.”
• “Architecture is complete once every layer has a PowerPoint colour and a package prefix.”
• “Put every decision in the service, every service in the base class, and every base class in a diagram.”

Biography

Frank trusts brokers because he has seen synchronous dependencies fail at inconvenient times. He is less enthusiastic about admitting when a queue is not the whole business process.

Typical worldview

Events create freedom when their contracts and consequences are explicit.

Favorite misconceptions

• “If the domain does not contain a topic name, nobody can prove the event really happened.”
• “A topic is already a workflow engine, provided nobody asks where the state went.”
• “A topic is the domain model after enough consumers depend on it.”

Biography

Mike values clear ownership and isolated change. His coffee gets cold while he explains why one more service will simplify the diagram.

Typical worldview

Independent deployment is the clearest form of architectural autonomy.

Favorite misconceptions

• “If `a().b().c()` compiles, the distributed design review has already happened.”
• “Split every step into a deployment and the architecture will coordinate itself through optimism.”
• “Every branch deserves its own deployable, preferably before lunch.”

Biography

Priya translates team autonomy into operable contracts. She likes guardrails, stable telemetry, and knowing exactly who owns the terminal state.

Typical worldview

Variation is healthy only when operations can still see and support it.

Favorite misconceptions

• “If both produce a JAR, the difference is clearly a YAML formatting preference.”
• “Move the broker first; meaning can follow during the incident.”
• “Two runtime concerns cannot coexist; one must win a ceremonial dependency-resolution duel.”

Biography

Nigel has a slide for every migration and a maturity model for every slide. He is useful precisely because somebody must ask whether the change is worth its ceremony.

Typical worldview

The right abstraction is usually one workshop away.

Favorite misconceptions

• “Add a fluent builder, a diagram, and a license tier for arrows; then it is strategic.”
• “Standardize every problem until the exceptions require a transformation programme.”
• “No change is safe because the application has survived too much to be questioned.”

Biography

Barry cares about compiler feedback, reproducible generation, and failures that happen before production. He distrusts magic that cannot explain its inputs.

Typical worldview

If the build accepted it, the result should be deterministic and diagnosable.

Favorite misconceptions

• “More generated files mean less evidence.”
• “Production is the best compiler.”
• “Every developer needs an annotation-processor minor.”

Biography

Carl treats generated artifacts as contracts, not implementation debris. He likes deterministic output and very specific diagnostics.

Typical worldview

Generated behavior is still product behavior and must remain inspectable.

Favorite misconceptions

• “Generate enough classes and the business will become decoupled by exhaustion.”
• “A stack trace is a user guide.”
• “The compiler now owns production because it wrote Java.”

Biography

Sam has shipped large Spring systems and does not confuse novelty with progress. He asks whether TPF complements the application or merely renames it.

Typical worldview

Existing application structure deserves evidence before replacement.

Favorite misconceptions

• “Put `@Pipeline` on a `@Service`, add three starters, and the architectural debate has been successfully dependency-injected.”
• “Business logic belongs wherever the injected `KafkaTemplate` is closest.”
• “Inject the producer into every service so no event feels lonely.”

Biography

Diego wants aggregates, events, and boundaries to carry precise meaning. He becomes suspicious when infrastructure vocabulary starts making business decisions.

Typical worldview

Language and ownership should reveal the business model.

Favorite misconceptions

• “If the method has a good aggregate name, infrastructure concerns can no longer enter the room.”
• “Every setter deserves a past-tense event, preferably before the entity finishes changing.”
• “No decision may execute until its aggregate has attended strategic-design training.”

Biography

Fran prefers values, immutable decisions, and visible effect boundaries. She also knows that purity without a practical shell is just a very elegant delay.

Typical worldview

Pure transformations are easier to reason about when effects remain explicit.

Favorite misconceptions

• “Compose the functions, call it a monad, and operational failure will politely become an implementation detail.”
• “If the function is pure, a customer will eventually materialize somewhere out of philosophical respect.”
• “Rewrite the whole core in one weekend; infrastructure will politely wait.”

Biography

Kai thinks in schedulers, containers, and blast radii. He asks where work actually runs and who notices when it stops.

Typical worldview

Topology is an operational choice, not a box somebody drew around code.

Favorite misconceptions

• “The I/O boundary is the cluster. Everything inside the cluster is technically a pure function.”
• “Portability means never reading the deployment docs.”
• “Every workload is a function after a timeout.”

Biography

Ada enjoys useful uncertainty but dislikes invisible authority. She wants AI calls treated as explicit external effects with typed inputs and observable outcomes.

Typical worldview

Models need narrower boundaries and better evidence, not larger prompts.

Favorite misconceptions

• “Ask a model whether the two implementations feel semantically similar.”
• “Ask a model for ubiquitous language; accept its first twelve nouns.”
• “The model has read the repository spiritually.”

Biography

Fred wants the framework to remove repetition without absorbing the application. He is happiest when customization remains supported rather than secret.

Typical worldview

A framework earns trust through coherent defaults and honest escape hatches.

Favorite misconceptions

• “If a getter does not emit telemetry metadata, it is probably hiding from its destiny.”
• “Hide all errors behind a generic failure.”
• “A rule is not real without a container.”

Biography

Jane has rescued enough hand-written persistence layers to distrust fashionable contempt for JPA. She asks whether data ownership and transaction behavior are actually clearer.

Typical worldview

Persistence is a boundary to model, not a historical embarrassment.

Favorite misconceptions

• “The transaction script is the domain model because it has the most annotations.”
• “A lazy proxy is a domain concept because it appears at runtime.”
• “If an event appears, immediately add a mutable entity and a cascade.”

Biography

Rita cares about idempotency, terminal outcomes, and whether another attempt is safe. She has learned that persistence is not the same thing as recovery.

Typical worldview

Retries are business decisions wearing runtime clothing.

Favorite misconceptions

• “No response means it definitely failed.”
• “Charge it again; accounting enjoys surprises.”
• “TCP was reliable, so the business outcome is known.”

Biography

Andy likes reactive systems when their scheduling facts remain visible. He does not treat non-blocking code as a personality or a moral achievement.

Typical worldview

Concurrency and backpressure are promises that require explicit ownership.

Favorite misconceptions

• “Blocking is a moral failure, not a scheduling fact.”
• “More threads are a substitute for backpressure.”
• “Fan out first; capacity will introduce itself.”

Biography

Helen uses boundaries to protect the model from transport and storage details. She objects when an abstraction hides which side effect actually happened.

Typical worldview

The domain should describe intent while adapters describe reality.

Favorite misconceptions

• “Add two interfaces and a factory; the side effect has now been ethically sourced.”
• “Rename the adapter `PortAdapterPort`, then no production failure can cross the boundary.”
• “Two interfaces and a factory solve every external boundary.”

Biography

Molly values quick feedback and controlled uncertainty. She asks whether a mock represents the real failure modes or merely agrees with the test.

Typical worldview

Fast tests matter, but doubles must not erase the behavior under test.

Favorite misconceptions

• “Mock the domain until it agrees.”
• “Mocks can represent a lost payment response.”
• “A stubbed 200 proves the partner integration.”

Biography

Terry balances unit speed with integration evidence. He distrusts both test pyramids used as dogma and production traffic used as acceptance testing.

Typical worldview

Use the smallest test that can prove the relevant contract.

Favorite misconceptions

• “Parallel systems are dangerous, so production should be the first and only rehearsal.”
• “If the new unit test passes, historical behavior has politely updated itself.”
• “Production traffic is the most realistic acceptance environment.”

Biography

Wendy has seen informal workflows become polling loops and correlation tables. She asks whether a pipeline is a typed flow or an orchestration graph in disguise.

Typical worldview

Coordination deserves durable semantics when the business truly needs it.

Favorite misconceptions

• “If something waits, retries, or has arrows, it needs BPMN and a committee trained to move diamonds.”
• “A compensation is a backward arrow travelling in a different colour.”
• “A timer is a business process.”