Graph, Spec, And Running Artifact Direction

This note records the current NEXUS direction for how application and system source truth may evolve over time.

Current Short Answer

Yes, that direction is broadly correct.

The likely long-term shape is:

  1. deeper semantic substrate
  2. derived spec or lens surfaces
  3. FORGE-style compilation or generation
  4. running artifact

But we are not fully there yet.

Current Refinement

The important refinement is:

  • the graph is a likely deeper semantic substrate
  • but not every concern should be forced into one graph prematurely
  • durable notes, path definitions, bounded contexts, and explicit lenses still matter
  • spec surfaces will likely sit between the graph and the running artifact

So the direction is not simply:

  • graph in
  • app out

It is more like:

  • durable semantics
  • graph-backed structural substrate where appropriate
  • lens- or domain-specific spec projection
  • FORGE transforms
  • running artifact

For Application Lines Like LaundryLog

The likely future shape is:

  • business meaning becomes durable and explicit
  • graph structure can carry important relationships and state transitions
  • selected projections produce reviewable specs
  • FORGE turns those specs into code and later into running artifacts

Examples of possible specs:

  • UI/view contracts
  • path definitions
  • event and command structures
  • component or token specifications
  • deployment or integration specifications

For NEXUS And FnHCI Lines

The same overall direction should eventually apply more broadly too.

That means:

  • NEXUS itself should become more modelable in its own terms
  • FnHCI and related interaction lines should also become more modelable
  • reusable libraries and app lines should be able to ride the same broader pattern

But this does not mean all of those lines are equally mature today.

What Is Still Missing

We are not fully there yet because several layers are still emerging:

  • stable graph ontology for the relevant domains
  • stable spec surfaces
  • explicit FORGE compilation boundaries
  • deterministic transforms from model to artifact
  • enough proven examples across more than one app or tool line

Current Working Rule

For now:

  • keep the durable business meaning explicit
  • keep the graph as an important emerging substrate, not a forced total abstraction
  • keep specs reviewable
  • keep generated or compiled artifacts downstream from those durable meanings

That keeps the system honest while still moving toward the longer-term graph/spec/artifact direction.

For the current clarification that Event Modeling is one lens over a fuller model rather than the whole model, see ATLAS, Lenses, And The Fuller Model.