The Identity of an Ecosystem

A valley before its records

The Cedar Valley Outcome Network is a teaching ecosystem with eleven principal referents. Anika is a grower and soil observer who works independently for most of the story. Cedar Farm Cooperative operates North Field, a cultivated field on land the cooperative does not own, and the Cooperative Grain Store where harvests wait for buyers. Anika belongs to Cedar Valley Growers, a community of growers that is neither the cooperative nor any single member. North Field drains into the Cedar Creek watershed, a living system that crosses two counties. Soil Probe 17 measures moisture in North Field. Harvest Lot A is the grain the field produced in its first season. The Grower Support Agent is software that drafts irrigation schedules for the cooperative under delegated authority. And the Water Stewardship Formation is a multi-party effort among the growers' community, the cooperative, and the watershed council to pursue one outcome: reduced irrigation withdrawals from the creek.

Notice what the list does not begin with. No paper, dataset, or web page appears among the eleven, because none of them is what the valley is about. Documents will appear — a moisture dataset, a permit, an impact report, an assessment — but they enter the story as evidence about the cast, and they are themselves a single family of referents among eight. The Seednote on the referent defines the term this Learning Center uses for all of them: the thing an RID identifies, as distinct from its names, descriptions, records, and representations.

Every edge in that figure is a hypothetical assertion that some party would make, not a registered predicate and not a fact the identity layer vouches for. The diagram is a picture of the ecosystem's topology as several organizations might describe it: who participates in what, what operates what, what lies within what, and what observes what. RSM ID's contribution is narrower and more durable than any of those edges. It gives each node one persistent identifier, so that when the watershed council writes about Cedar Creek and the cooperative writes about North Field and the soil observatory writes about Soil Probe 17, each of them is unambiguously writing about the same thing the others mean.

Eleven identities, none of them derived

Each of the eleven referents receives its own RID, and each RID is opaque. Anika's illustrative RID is 451.9J0RFXDGKFYE4CXS; Cedar Farm Cooperative's is 451.WMW78429ZAZXP882; North Field's is 451.8DKZW0Q7EV8WW3VQ; Cedar Creek's is 451.4ZFNMDBX5YHB833N. None of them encodes a type, an owner, a county, or the space that first described the referent, and none of them is derived from another. That last point is the one most easily lost. Anika's identity does not come from her membership in the Growers or from her season working for the cooperative. North Field's identity does not come from the cooperative that operates it, and Cedar Creek's does not come from the council that studies it. Identity in RSM ID is not inherited down a chain of affiliations; it attaches directly to each referent.

That independence is what lets the graph survive change. If North Field's identity were a sub-identifier of the cooperative, a change of operator would leave the field identified by the wrong organization. If Anika's identity were issued as a staff record, leaving the cooperative would end it. The Fieldnote on why a person does not belong to a space follows that argument for people, and A farm is not its owner follows it for land.

Independence is not the same as unregulated issuance. Every RID is still issued by a registered authority under a delegated prefix, with atomic registration and auditable provenance, and the issuance authority that registers a referent acquires no ownership of it. What the architecture rules out is a dependency of the referent's identity on any particular organization, membership, or Semantic Space. What it does not yet provide is a finished operational pathway for every kind of referent, and this Learning Center marks those gaps rather than filling them with invented mechanisms.

Three vocabularies and a registry gap

When the cooperative describes North Field, it may classify it as rsm:Place. When the registry names it, the name rrn:451:cedarvalley:us-ca:cedarcoop:place/north-field uses the registered kind place. And when this Learning Center groups it with Cedar Creek under "places and living systems", it is using a teaching family. Those are three different vocabularies, and the ecosystem exposes exactly where they diverge. Cedar Valley Growers is described as rsm:Community, Harvest Lot A as rsm:Material, and the Grower Support Agent as rsm:DigitalAgent, yet none of those has a registered naming kind in the minimum profile. The divergence also runs the other way: Soil Probe 17 can be named with the registered kind instrument, but Document 08 registers that kind without defining it, while the RSM model uses rsm:Instrument for legal, financial, measurement, and other constructs, so which semantic class describes a physical sensor is itself unresolved (SPEC-10).

The consequence is precise. Because an RID encodes no kind, nothing in the RID grammar prevents the Growers or the grain lot from having one; the illustrative RIDs 451.358VPJMD1B17MKZ4 for the Growers and 451.K9PSWTQ7MSXE7VN6 for Harvest Lot A are as well-formed as any other. What is missing is a registered way to name them, and possibly a rule about what kind their Identity Records may carry. That is an open registry question, recorded as SPEC-03, and the corpus shows no RRN for either referent rather than inventing a community or material kind that would be rejected at authoritative registration. The Seednote on community and the Viewgraph What RSM ID identifies set out the distinction.

Many spaces, one watershed

Eight independently governed Semantic Spaces describe the Cedar Valley cast. The cooperative's operations space records field operations, inputs, irrigation, and storage. The watershed council's space records streamflow, water quality, and ecological condition. A county water office keeps permits, allocations, and jurisdictional interpretations. A university soil observatory keeps soil observations and calibration records. A grain buyer keeps contracts and lot records. Anika keeps a personal space for her own observations and the claims she chooses to disclose, which Document 08 permits because a person may administer Semantic Spaces of their own (§4.6). The formation keeps a coordination space, and an independent assessor publishes signed assessments of outcome evidence.

Read the arrows carefully. Each points from a space to a referent, and each is an account that the space maintains under its own governance. None of the spaces contains Cedar Creek. The watershed is not inside the council's database, the water office's permit system, or the cooperative's farm-management software; it exists in the world, and each of them holds knowledge about it, refers to it by its RID, and decides who may see what it holds. That is the distinction drawn in the Viewgraph Identity plane and knowledge plane: the registry knows which watershed, the spaces know about it, and the watershed itself belongs to neither plane.

The practical benefit is that these spaces do not need to agree on a schema, a database, or an ontology in order to cooperate. They need only agree on which referent each statement concerns. Proposed infrastructure such as Locus would serve each space's knowledge, and Atlas would let a reader traverse from the council's account of the creek to the cooperative's account of North Field, keeping every assertion attributed to its source.

Change without erasure

The ecosystem is followed from 2031 to 2034, and almost everything in it changes. In 2031 Anika works a season for the cooperative and then leaves to observe soils independently; one relationship ends and her RID is untouched. The same year Soil Probe 17 is recalibrated and later redeployed to another field; calibration and deployment are events with their own records, and the device remains the same device. In 2032 the cooperative renames itself and registers the successor name rrn:451:cedarvalley:us-ca:cedarcoop:organization/cedar-valley-farm-cooperative, while its original name stays reserved and keeps resolving to the same organization. In 2033 stewardship of Cedar Creek passes from the council to a joint authority, which changes governance metadata and leaves the watershed's identity alone, exactly as Document 08's watershed example describes.

Two events in the timeline are deliberately harder. In 2032 North Field is resurveyed and its recorded boundary moves. Whether that is a refined description of the same field or the beginning of a different referent depends on the continuity criteria governance sets for fields, and a small correction plainly differs from a subdivision into two fields. In 2033 Harvest Lot A is divided between two buyers, and nothing in Document 08 says whether the two parts continue the lot's identity, become new referents with a recorded lineage, or both. This Learning Center treats both as open continuity questions (SPEC-04) rather than declaring a universal rule. The lifecycle states of Document 08 §18.4, including superseded for a distinct successor referent, provide the vocabulary for recording a decision once governance makes one; they do not make the decision.

Evidence, disagreement, and privacy

In 2033 the formation's coordination space publishes an impact report asserting that irrigation withdrawals fell by the targeted amount, citing meter records and the moisture dataset from Soil Probe 17. In 2034 the county water office challenges the figure, asserting a smaller reduction on the basis of its own allocation records. Both assertions are about the same outcome, identified as 451.FVB2FXJAH0YQ87GN for reduced irrigation withdrawals. Sharing that RID establishes that the two parties are disagreeing about one thing rather than talking past each other; it does not establish which of them is right, and the identity layer has no authority to decide. The Seednote on claim and evidence and the Fieldnote An outcome is not a claim separate the target, the measurement, the attestation, and the verified result. How such disputes are mediated across spaces is open (SPEC-06).

Evidence follows the same discipline. A reading from Soil Probe 17 is an observation about North Field, recorded by the observatory, and it supports or weakens an assertion; it is not the probe, not the field, and not proof that the outcome occurred. The Fieldnote on sensors and readings draws those lines in detail.

Disclosure differs by referent as well. A public request to resolve Cedar Creek may return its identity status, the spaces that publish descriptions of it, and a public landing page. The same request for Anika's identity returns, to anyone not authorized, the response a nonexistent identity would produce, so that the resolver cannot be used to discover whether she has an RID at all. Both outcomes are policy decisions about representations; neither alters which referent each RID identifies. The Fieldnote on identifying without exposing explains why enumeration resistance matters most for people.

Five separations, and a counterexample

The Cedar Valley story is a demonstration of five separations that hold across every family of referents. A referent is not its representation: Cedar Creek is not the council's map of it, and Anika is not her profile. Identity is not relationship: Anika's RID exists before, during, and after her affiliations. Issuance authority is not ownership: the namespace authority that registers North Field gains no title to it, consistent with Document 08's rule that issuance confers no legal ownership, consent, access rights, or authority to disclose (§8.5). The registry is not a Semantic Space: the identity record holds lifecycle, names, and bindings, while knowledge stays with the spaces. And semantic equivalence is not identity equality: two records describing similar fields are not thereby the same field.

The counterexample shows why the last separation matters. Suppose the grain buyer's system and the cooperative's operations space each hold a record of a field named "North Field, Cedar Valley", with overlapping boundaries and the same operator. An aggregator tidying its search results merges them. But the buyer's record actually describes a neighbouring parcel that the cooperative also called North Field before a resurvey. Every contract, observation, and permit attached to the merged record now points at a field that does not exist, and the error is invisible because the merged view looks cleaner than the originals. RSM ID prevents that outcome by refusing to infer identity from resemblance: where two systems may have represented the same entity, reconciliation keeps evidence and requires an authorized decision (§15.3).

Conclusion: identify the constituents, let the spaces describe them

Following one valley for four years shows what a common identity layer adds to an ecosystem and what it deliberately leaves alone. It adds one persistent, opaque identifier for each constituent — a person, an organization, a community, a field, a watershed, a facility, a sensor, a grain lot, an agent, a formation, and an outcome — so that independent accounts can be joined without guessing. It leaves to the Semantic Spaces everything they are responsible for: descriptions, observations, permits, contracts, assessments, and their disagreements, each governed by whoever holds it. And it keeps the two apart, so that a rename, a new steward, a recalibration, or a dispute changes the right records and leaves identity intact.

For practitioners, the implications are concrete. Identify the referents your systems actually care about before identifying the documents about them. Record affiliations, operation, custody, and stewardship as dated relationships with their sources rather than as parts of an identifier. Refer to other parties' referents by RID and keep your assertions attributed. Treat splits, merges, and boundary changes as governance decisions that need criteria, not as automatic outcomes. And never treat a shared identifier as agreement.

The ten-scene Journey of an RSM Ecosystem tells this story at a glance. From here, the six canonical journeys follow each family in depth: A Person Before an Affiliation for Anika, Place Is Not a Coordinate for North Field and Cedar Creek, and The Identity of a Formation for the water effort and its contested outcome. For the reasoning behind persistent identity itself, return to Why Persistent Identity Matters.