Schema Draft¶
Document — instantiates Cognitive Representation Externalization
A provisional document defining categories, fields, relations, and constraints in a model.
A Schema Draft writes down, in a semi-formal document, the categories a domain contains, the fields each carries, the relations that connect them, and the constraints that must hold — but with the word draft doing real work. Its defining move is to externalize a mental model at the rigor of a type system while explicitly holding that type system open for revision: it commits to precise category boundaries and named relations (one-to-many, must-have, mutually-exclusive) so that hidden classification assumptions become checkable, yet it is authored as a provisional document whose whole purpose is to be marked up, contested, and rewritten before anything is built on it. That combination — formal enough to catch a contradiction, unfinished enough to change cheaply — is what makes it neither a sketch nor a finished specification.
Example¶
A regional museum is building a catalog and its curators keep colliding over what an "object" even is. One models a three-panel altarpiece as one object; another as three. A cataloger drafts a schema: entities Object, Component, Maker, Provenance Event; relations Object has-many Component, Object attributed-to Maker, Provenance Event transfers Object; constraints every Object must have at least one Provenance Event; a Maker attribution may be uncertain. Writing it down formally forces the altarpiece question into the open: the has-many Component relation resolves it — one Object, three Components — a decision that was invisible while it lived in two curators' heads as an unspoken assumption. The draft also carries an explicit boundary note: "digital-born media are out of scope for v1; loans are modeled as Provenance Events, which may be wrong." Circulated for review, it comes back with red ink on the uncertain attribution rule, and the cataloger revises the schema before a single record is entered.
How it works¶
- Name the categories and their fields. Entities are given crisp definitions and the attributes each must carry, which forces the fuzzy "kinds of things" in people's heads to acquire edges.
- Encode the relations and constraints. Links are typed (one-to-many, required, exclusive) and rules are stated, so the model can be checked for contradiction rather than merely read.
- Mark the scope and the shaky assumptions. What the schema deliberately does not model, and which choices are provisional, are written into the document itself.
- Circulate, redline, and re-issue. The draft exists to be revised: reviewers mark disagreements, and the schema is re-versioned until it stops drawing objections — then it may graduate out of "draft."
Tuning parameters¶
- Formality — prose definitions versus a strict, machine-checkable notation. More formality catches more contradictions but raises authoring cost and can exclude non-technical reviewers.
- Constraint strictness — loose guidance versus hard rules with cardinalities. Strict constraints catch invalid states early but risk over-committing before the domain is understood.
- Scope breadth — a narrow core schema versus a comprehensive one. Narrow drafts stay revisable; broad ones try to settle everything at once and freeze prematurely.
- Provisionality signaling — how loudly the "this is a draft" and open-question markers are flagged. Strong signaling keeps it revisable; weak signaling lets a draft harden into an unexamined standard.
When it helps, and when it misleads¶
Its strength is forcing classification assumptions into the open at a rigor where contradictions actually surface — it is the mechanism for the moment a team needs typed categories and constraints, not just a loose sketch, and it catches the "we meant different things by this term" failure before it is baked into a system.
Its failure mode is premature freeze: a schema's very formality lends it authority, so a draft can calcify into a de-facto standard nobody dares revise, encoding an early, wrong classification into everything downstream. The classic misuse is treating the first draft as final because it looks rigorous. The guarding discipline is to keep the "draft" and the open-question markers loud, to date and re-issue versions, and to leave the schema formally provisional until it has survived review — the opposite of the informal, free-text Concept Map, which stays deliberately un-typed because its job is to generate rather than to specify.[1]
How it implements the components¶
relation_encoding— its core content is typed relations and constraints (one-to-many, required, exclusive) that make the links among categories precise and checkable.representation_boundary_note— the explicit scope statement and provisional-assumption markers record what the schema omits and where it may be wrong.revision_loop— authored as a draft to be redlined and re-issued, it carries the update pathway by which the model is corrected before it is trusted.
It specifies structure rather than eliciting it — it does not draw an unformed model out of someone's head (implicit_model) — that is Canvas Template — and it does not itself run the inspection that surfaces objections (shared_review, mismatch_capture), which is Model Walkthrough Review.
Related¶
- Instantiates: Cognitive Representation Externalization — the Schema Draft externalizes classification and relational assumptions at a formal, revisable rigor.
- Sibling mechanisms: Canvas Template · Comparison Table · External Memory Aid · Model Walkthrough Review · Process Flowchart · Whiteboard Modeling Session · Concept Map · Causal Loop Diagram
Editorial Notes¶
Form Classification¶
Form family: Representation, Specification & Plan
Rationale: Schema Draft operates as a static representation, map, specification, schema, or prospective plan that externalizes information because it a provisional document defining categories, fields, relations, and constraints in a model.
Independent corroboration: The frozen evidence defines Schema Draft as 'A provisional document defining categories, fields, relations, and constraints in a model', so its operative form is Representation, Specification & Plan.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Library & Information Science
Origin pattern: Convergent development
Present-day reach: Universal
Rationale: Provisional definition of categories, fields, relations, and constraints is a knowledge-organization artifact.
Related originating lineages:
- Computer Science & Software Engineering — Computer science and software-engineering practice supplies a parallel or contributing lineage for the mechanism's defining operation: a provisional document defining categories, fields, relations, and constraints in a model.
- Mathematics — Formal structure materially supplies relations and constraints.
Review resolution: Both blind reviewers agree that library_information_science is the primary historical origin. Explicit reconciliation of alternate_origin_disagreement, origin_mode_disagreement, domain_reach_disagreement, encyclopedia_synthesis_disagreement starts from reviewer_a's mechanism-specific evidence: Provisional definition of categories, fields, relations, and constraints is a knowledge-organization artifact. Reviewer A proposed alternates=computer_science, mathematics, origin_mode=convergent, domain_reach=universal, and encyclopedia_synthesis=true; reviewer B proposed alternates=computer_science, origin_mode=single_lineage, domain_reach=multi_domain, and encyclopedia_synthesis=false. The final record retains every independently supported alternate from either review (computer_science, mathematics) without an arbitrary cap, selects origin_mode=convergent to represent the combined lineage evidence, and records domain_reach=universal and encyclopedia_synthesis=true. Present-day transfer is recorded as reach and is not treated as proof of historical origin.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Reconciled after independent review; high confidence.
Notes¶
A Schema Draft is often the formalization of what a looser mechanism surfaced first: the entities and relations a Concept Map or Whiteboard Modeling Session generated informally can be promoted into a schema's typed structure once they stabilize. Keeping it explicitly a draft is what preserves the revisability the archetype demands.
References¶
[1] The entity-relationship model (Peter Chen, "The Entity-Relationship Model — Toward a Unified View of Data," 1976) is the canonical notation for stating a domain's entities, relationships, and constraints formally — the rigor a schema draft reaches for while keeping the result open to revision. registry ↩