Decisions in the graph

Why is it built this way? ADR-style decisions wired to what they replaced, produced and touched.

Architecture decision records usually rot in a folder nobody opens. In Arkaik, a decision is a node in the same graph as your screens and endpoints, so the "why" sits right next to the "what".

Each decision has its own status (proposed, approved, enacted, and the endings: rejected, deprecated, superseded) plus the date it was actually decided. Its lifecycle status follows along, so an enacted decision reads as live and a superseded one as archived.

Three connections do the heavy lifting. Supersedes links a decision to the one it replaces, and flips the old one to superseded. Generates links a decision to the acceptances it produced. Impacts links it to the flows, views, data models or endpoints it changed.

So when someone asks "why does onboarding skip the email step?", you don't dig through chat history. You click the screen and follow the edge.