Keep pulling the thread on Neal Ford & Mark Richards.
Neil Ford and Mark Richards are developing a concept called a "governance mesh," designed to operate within a service mesh to manage and run architecture fitness functions.
Neil Ford and Mark Richards are developing a lightweight architecture definition language (ADL) to serve as a platform-agnostic pseudocode for specifying architectural constraints.
The lightweight architecture definition language (ADL) is designed to be used with Large Language Models to generate executable fitness functions in languages such as Java, .NET, Python, and TypeScript.
The Architecture Definition Language (ADL) is being designed to be "prompt friendly" for use with LLMs like ChatGPT to automatically generate executable tests for frameworks such as ARC unit and net arc test.
Neil Ford and Mark Richards will be conducting workshops on their "Architecture is Code" material at conferences in Europe during May and June.
Unlike unit tests, most architecture fitness functions are not written in the application's primary language but are often implemented as monitors, GitHub Actions, or scripts in languages like Ruby.
Tools such as PlantUML, Structurizer for the C4 model, and Archimate can export diagrams as source code, but this output describes the architecture without being executable to enforce its constraints.
The lightweight architecture definition language (ADL) developed by Neil Ford and Mark Richards will be featured in their upcoming book titled "Architecture is Code".
For a "repo per service" architecture, a fitness function can be implemented as a GitHub Action that detects copy-pasted code to alert architects when duplication exceeds a defined threshold.