node flow
Nodes are primarily constituted by their node flows, which are evolvable datasets about their node's metadata, configuration, referent, or payload datasets. They exist through time, independent of any specific version or realization, and can evolve semi-independently from other flows
There are five types of node flows.
- metadata flow (required)
- operational config flow (optional)
- inheritable config flow (optional)
- reference flow (optional)
- data flow (for data nodes)
Relationship to snapshots
As DatasetSeries, node flows are realized through flow snapshot datasets, which are temporal slices of the flow. To borrow a phrase from the PROV model, we say that a snapshot is a specialization of the node flow.
Relationship pattern:
Every node flow has at least two concrete snapshots: current snapshot and next snapshot.
The node flow is a DatasetSeries and may have multiple version snapshots.
Ontology Example
- node flow: "My ontology definitions" (persistent concept)
- flow snapshots: v1, v2, current version, working draft of next version (specific realizations)
/my-ontology/
└── _data-flow/ ← node flow (ontology definitions)
├── _current/ ← flow snapshot (in this case, probably )
├── _next/ ← flow snapshot
├── _v1/ ← flow snapshot
└── _v2/ ← flow snapshot
In this example:
Each _current/, _v1/, etc. contains flow snapshot realizations
Persistent Identity
node flows provide conceptual continuity by:
- Maintaining meaning across versions and changes
- Preserving references from external sources
- Enabling evolution while keeping identity stable
- Supporting versioning without losing conceptual coherence
Children
Backlinks