Kafka Frank
“If the domain does not contain a topic name, nobody can prove the event really happened.”Join the conversation
The conference break that keeps going
The Coffee Machine
Real architecture questions, recognizable engineering instincts, and answers that admit the trade-offs.
Search the conversationsBrowse
By topic
Why TPF ExistsThe architectural argument for strongly typed application flows.8 questionsBring Your Existing AppAdopt TPF without pretending your Spring, JPA, or Kafka estate disappeared.8 questionsDomain ModellingBoundaries, events, aggregates, sagas, and useful domain language.8 questionsRuntimeReactive execution, contracts, diagnostics, and runtime behavior.6 questionsConnectorsTyped I/O boundaries and the realities they admit or publish.3 questionsDeploymentRuntime placement, topology, containers, functions, and portability.7 questionsOperationsRetries, replay, failure ownership, and operational timelines.3 questionsTestingProving contracts and failure behavior without booting the planet.8 questionsAIMachine-readable contracts and safe AI boundaries.2 questionsGovernanceGuardrails, ownership, autonomy, compliance, and escape hatches.11 questions
Random quote
Fortune from the machine
fortune tpf
“I/O is not clearer because it is injected; it is clearer because someone owns its consequences.”
Today’s debate
What trade-off is this missing?
Fresh pot
Recently added FAQs
Domain ModellingAre domain events first-class, or merely outputs from a pipeline?TestingAre generated artifacts stable enough to review in pull requests?Bring Your Existing AppCan a legacy monolith participate without being fully converted?TestingCan business rules be tested as ordinary Java functions?GovernanceCan developers customize generated code without losing regeneration safety?GovernanceCan different teams use different resilience policies without fragmenting operations?
Choose a chair