How this is modelled
This index is the record, not a mirror of someone else's. A quotation here is an entity with its own identifier, its own citation and its own history.
Entities
- Person
- A human with a stated relationship to North or South Carolina. A person does not need an entry anywhere else to exist here.
- Affiliation
- The reason a person is in the index, stated rather than implied. A relationship has a kind, and kinds are not interchangeable.
- Quotation
- Words a person said or wrote, carrying a record identifier of the form
CQ-XXXXXXXXthat survives every rebuild. - Attestation
- Where the words occurred: the speech, the book, the broadcast, with its year and locator. This is the citation a reader wants.
- Provenance
- How the record entered this index. Deliberately separate from the attestation, because where words were spoken and where an index read them are different facts.
- Portrait
- A likeness from the Library of Congress, used only where the Library states the item is public domain. A name match alone never earns one: agreeing life dates or a recorded occupation must corroborate it, and dates that disagree are a veto.
- Saying
- An expression belonging to the state rather than a person, kept apart from quotations so it is never dressed up as one.
- Submission
- What a reader sends in. It stays a submission for its whole life and becomes public only when a person publishes it.
State connections
Each affiliation records state_code: NC or SC. One person can have relationships with both states. A person's standing is derived separately for each state. The combined collection counts each person once.
For example, Dizzy Gillespie was born in Cheraw, South Carolina, and studied at the Laurinburg Institute in North Carolina. His profile carries both connections.
Books and library records
A person can connect to several editions and library catalogs. Each library record keeps its provider identifier, author match, publication year and access link. Quotation citations point separately to the edition and page where the words appear.
Standing
The correction at the centre of this model. A person's standing is the strongest relationship they hold.
- native born or raised in the named state
- resident lived here, or died here with another tie to corroborate it
- affiliated studied, taught, worked or held office here, only died here, or a town here is named for them
Richard Nixon took a law degree at Duke and Michel Chossudovsky took a degree at Chapel Hill. Both have a real, checkable North Carolina relationship. Neither is a North Carolinian, and a dictionary that calls them one is wrong about the very thing it exists to be right about.
Death needs corroboration for the same reason: hospitals and prisons are full of people who did not live in the state. Ted Kaczynski died at a hospital here and has no other tie to it. Randall Jarrell died in Chapel Hill and taught at two of its universities. A rule that cannot tell those two apart is not a rule.
Namesakes are a tie of their own. Walter Raleigh never set foot in North Carolina, yet its capital carries his name, and a Carolina dictionary without him would be wrong. The tie counts only when a settlement is named for someone: a foundation, a school or a ship named after them is an institution's choice, not the state naming itself.
Relations in use
| relation | records |
|---|---|
educated_at | 396 |
born_in | 212 |
employed_by | 149 |
died_in | 75 |
lived_in | 57 |
worked_in | 36 |
held_office | 17 |
namesake_of | 13 |
raised_in | 5 |
Rules the model enforces
- A person with no stated affiliation does not exist here.
- A quotation always resolves to exactly one person.
- A quotation always carries a provenance record, whatever its corpus.
- Standing is derived, never typed in, so it cannot drift from the relations.
- Words about a person are not words by them, and are never stored as theirs.
- Doubt is a property, not an omission: a contested attribution shows its status.