{
  "resourceType": "CodeSystem",
  "status": "draft",
  "content": "not-present",
  "name": "SeedIdCS",
  "id": "seed-id",
  "title": "Provisioned Resource Seed Id CodeSystem",
  "description": "The authoring id a provisioned resource carried in the bundle it was loaded from, kept so the\nresource can be found again after the store filed it under a different id. Carried as a tag on\n`meta.tag`, where this is the tag's `system` and the authoring id is the `code`.\n\n**Why it exists.** Tenant initialisation loads a fixed set of bundles into a tenant's store, and the\nid a resource is stored under is derived per tenant rather than taken from the bundle. The authoring\nid is therefore the only stable handle for \"the same provisioned resource, in whichever tenant\", and\nresolving through it is what makes a re-run adopt the copy already in the store instead of creating\na second one.\n\n**A tag, not an extension or an identifier.** The address identifies the `system` of a `Coding` in\n`meta.tag`, which is why it is declared as a code system. It is carried on `meta.tag` because an\nauthoring id is a label about where the resource came from rather than a name for what it\nrepresents: it identifies a row in a provisioning bundle, is meaningful only to the loader that\nwrote it, and identifies the patient or document to nobody. `identifier` would assert more than is\ntrue. (An earlier version of this page justified the choice by saying an identifier search would\nneed a custom `SearchParameter` registered first. That is not so -- R4 publishes an `identifier`\nsearch parameter for every resource type a loader touches -- and the reason above is the one that\nholds.)\n\n**Codes are not enumerated, and `content` is `not-present`.** A code is one resource's authoring id,\nminted with the bundle it lives in, so the membership of this system is the bundle corpus and\nchanges on that corpus's cadence rather than this specification's. Codes are opaque: do not parse\none, and do not infer a resource type, a tenant or an ordering from it. They are written in lower\ncase.\n\n**No code here is well known.** A deployment may probe for a single known authoring id to decide\nwhether a store has been provisioned yet; that id is an ordinary member of the corpus and carries no\nstanding in the vocabulary.\n\n**It is provenance and nothing else.** Presence of the tag says a resource arrived through tenant\ninitialisation and what it was called there. It asserts nothing about the resource's clinical\nmeaning, nothing about whether it may be edited or deleted, and nothing about who may read it.\nAbsence says only that the resource did not arrive that way.",
  "url": "https://fhir.apps.health/CodeSystem/seed-id",
  "experimental": false,
  "caseSensitive": true
}