Reduce the time between agreeing on a concept and running it
Measure: model-to-runtime time
The model writes the software.
A concept modeled once in AML carries its type, its definition, its constraints and its measures. That is enough for the generator to emit the aggregate, its events, its command handlers, its schema and its tool surface — which is how ageni's own apex product was built.
What it moves
Each outcome names the measure it moves and the direction it moves it — the value is stated as a measure, not as a claim.
Reduce the number of places one concept is described
Measure: duplicate descriptions per concept
Increase the share of a system derived from its model
Measure: derived surface share
3
desired outcomes
4
mechanisms
3
AML solutions
Hand-written code drifts away from the model that justified it.
A specification, a database schema, an API contract and a UI form are four descriptions of the same concept, each maintained by hand and each free to disagree. AML removes the copies: the concept is stated once as a typed model, and every surface is derived from it, so there is nothing left to keep in sync.
before
resolved
The genome type system holds the metaclasses a concept can register as, and the suffix rules that decide which. From a typed concept the emitter derives the value objects, the event-sourced aggregate root, the command and event pairs, the persistence schema and the exposed tool entry — each one a projection of the same model rather than an independent artifact.
- 1 A concept registers as a metaclass through its terminal suffix, so its manifestation is derived, not chosen by hand.
- 2 Every property descends to a data type through value objects — no raw primitives on a command, event or aggregate.
- 3 Commands derive from a root verb and emit the matching past-participle domain event.
- 4 The generated module ships its own schema, its own command handlers and its own tool entry.