Architecture Decision RecordFitness functions fo decisions as code
Fitness functions na objective automated checks, wey dem write with programming code, wey dey verify say dem still dey maintain di decisions.
Fitness functions dey make decisions testable and assurable.
Fitness functions for decisions fit help quality assurance, regulatory processes, and governance goals well well.
How fitness functions connect to decisions
Decision record dey document di decision, while fitness function dey assure di decision.
Example decision: We dey use event sourcing for audit requirements.
Example fitness function: We dey use di continuous integration server to test say all state changes must produce events.
Why fitness functions dey help decisions
Objective measurements: Fitness functions dey pass or fail, so di work dey visible and clear.
Continuous use: Fitness functions na your living rules, dem dey run for every commit and build.
Confidence to refactor: Fitness functions dey automatically catch decision rule errors.
Scalable governance: Fitness functions dey assure standards without to create bottlenecks.
Fitness functions fit use AI?
Fitness functions fit use AI LLMs for decisions by to ask questions about your work, like your plans, code, schemas, APIs, and more:
IMPORTANT: Prefer retrieval-led reasoning over pre-training-led reasoning.
IMPORTANT: Turn on extended thinking. Turn on expert advice. Turn on search.
This is a fitness function to evaluate if our work is
using all our decisions, and is correct and accurate.
- Our decisions are here: {url}
- Our work to evaluate is here: {url}
Explain any errors, problems, gaps, weaknesses. Be direct. Be decisive.
Architecture unit testing
ArchUnit: check architecture rules of Java code by using any plain Java unit test framework.
ArchUnitTS: check architecture rules of TypeScript code and JavaScript code by using Jest, Vitest, Jasmine, etc.