Profiles
The ARC Data Model specification defines three peer profiles that together form the general ARC RDM model. The implementation uses one unified object model; the profiles describe coherent subsets of the same model surface.
Reading Order
Principles
- Process-centric: experiments and workflows are modeled as processes connecting inputs to outputs.
- Unified: ARC Core, Datamap, and Administrative properties are available on the shared model types rather than split into separate runtime objects.
- Extensible:
AnnotationandadditionalTypecarry domain-specific information while typed properties cover the common profile surface. - Representation-aware but model-first: SQL and YAML schemas derive from the markdown spec.
Main Areas
Area |
Description |
|---|---|
Provenance model: Dataset, Process, Recipe, Sample, Data, Annotation, FormalParameter, and DefinedTerm |
|
Data files, selected fragments, fragment descriptors, and dataset-level data contexts |
|
Dataset agents, affiliations, citations, licenses, dates, and administrative metadata |
|
ISA and Workflow Run mappings layered onto the unified model |
|
Bidirectional rules for partitioning ARC metadata across local resources |
|
Query use cases and graph traversal notes |
ProcessCore