Story architecture glossary

Zero drafting is the practice of writing a scene at speed, against a visible record of what the story world has already established, for the purpose of discovering what that world cannot yet support. The pages are not the deliverable and are not kept. The deliverable is the list of decisions the world turns out to still owe you.

Distinct from the loose usage in which a "zero draft" simply means a messy first pass. That usage grades the draft; this one disposes of it. A zero draft is never revised into the finished work — it runs alongside and ends in deletion by design.

The set of facts that are true inside a story world. Canon is not everything an author has imagined and not everything that appears in a draft — it is specifically what the work treats as established and will be held to later.

The practical test is whether contradicting it would be an error rather than a choice. Anything that fails that test is still a possibility, not canon.

The gradual divergence between what a story world has established and what its creators believe it has established. Drift is caused by ordinary forward motion — decisions made months apart, each individually sound — rather than by carelessness, which is why it survives careful work.

One authoritative record that every collaborator checks rather than remembers. It is the precondition for every other continuity practice: a record that is out of date, or that competes with two other documents, is worse than none, because it produces confident wrong answers instead of honest uncertainty.

Two or more incompatible versions of a story world held simultaneously by different collaborators, each believing theirs is the agreed one. It is the characteristic failure of writers' rooms and distributed production teams, because verbal consensus in a break produces as many reconstructions as there were people present.

A continuity failure in which a character acts on information they could not have had at that point in the story. It is the most common continuity error in produced television and the least visible in an outline, because summary does not track information as it moves between characters.

Ages, dates, durations, and distances that fail to reconcile once written out as a single sequence. Each number is usually correct in isolation, which is precisely why the error survives review — it only becomes visible when a scene forces several of the facts to be true at the same moment.

Two established facts that cannot both be true. The simplest continuity failure to describe and among the hardest to notice, because a contradiction in prose is syntactically perfect — both statements are grammatical, readable, and survive every automated check that exists.

The scene in which the largest number of separately established elements have to be true simultaneously — characters with distinct histories, a world rule, a location, and a fixed point on the timeline, all loaded into one room. It is the correct target for a zero draft, because load concentrates at intersections.

The two lists a zero draft produces. Missing holds things the world never decided; Broken holds things the world did decide that the scene needs to be otherwise. Missing items are work. Broken items are contradictions caught at the cheapest possible moment, and each one is a fork: either the scene is wrong or the canon is wrong.

An animation-specific second pass that lists every spatial and physical fact a scene silently assumes — positions, distances, sightlines, what characters hold and touch — and checks that list against model sheets, prop designs, and approved boards. It catches errors that read as fine in a script and become impossible in a drawing.

A deliberate change to established canon. A retcon differs from a plot hole in intent and in bookkeeping: it is decided once, written into the record, and its consequences are traced. An accidental contradiction is neither decided nor recorded, which is what makes it expensive.

A review pass over existing pages that checks them against the established record. It is the last cheap moment to catch an error rather than the first, because by definition the pages already exist — which is why it pairs with a practice that runs before anything is written.

The document that holds a production's canon — characters, world rules, geography, timeline, and the decisions a team is expected to work from. Its value is entirely a function of whether it is current and authoritative; a bible people no longer trust is a bible people stop reading.

A constraint the story world obeys — how the magic works, what the technology cannot do, what a power costs. World rules are load-bearing in a way most canon is not: breaking one does not merely contradict a fact, it retroactively weakens every scene that respected it.

The internal consistency of a story world considered as a system rather than as a list of facts — whether its rules, histories, and relationships can all hold at once. CanonBoard describes itself as a world logic engine because the object it operates on is the system, not the prose.

A systematic scan across an entire story world that surfaces contradictions no single reader happened to notice. It is the one operation where a tool genuinely beats a person: the number of fact pairs that can conflict grows as the square of the facts, and human working memory does not grow at all.

In CanonBoard this reads your canvas, not your draft. It can only check what has been written down — which is why it complements zero drafting rather than replacing it.

A contradiction or unexplained gap that a reader or viewer notices. Worth distinguishing from a continuity error in general: every plot hole is a continuity failure that escaped, but most continuity failures are caught before anyone outside the production ever encounters them.

Audience funding is the practice of financing a story project in small, repeated amounts from the readers already following its development, rather than in a single campaign for a finished product. The money attaches to a world that is still being built, and at its most specific to individual decisions inside it. The work is therefore paid for while it can still change, and the people paying can see exactly what their contribution is attached to.

Distinct from crowdfunding, which is a campaign: a goal, a deadline, reward tiers, and a finished product promised in advance. Audience funding promises nothing and has no end date, because the money is attached to work still being made rather than to an object being sold.

A story project published as a public page, where an audience can read the world, follow it, discuss it, vote, and submit pitches for what should happen next. Publishing a Board is also what makes a project able to receive funding — an unpublished project cannot take money.

Publishing a Board does not publish everything. Cards held in the Locked Vault stay private, so material a creator is building toward remains invisible on the public page until they release it.

Putting money behind one specific submitted idea on a public Board, to raise its visibility with the creator. Backing pays the creator of the project, never the person who submitted the pitch, and it does not place the idea in the world's canon.

The common misreading is that a backed pitch is binding. It is not — canonization remains entirely the creator's decision. A backed pitch is best understood as a demand signal, indicating which direction of a world an audience finds compelling enough to pay for.

The act of a creator accepting a submitted pitch into the established canon of their world. Canonization is always the creator's decision, can be declined for any reason, and confers credit on the contributor rather than compensation, ownership, or any other claim.