Chapter 10. Design Elements and Working Principles

Now that the parts are known, we move on to process. The first chapter of Part 3, Methodology, draws a map of the design work as a whole. What is a world made of? Where does the work begin, and what is its nature as it proceeds? What must it finally deliver? The guiding question of this chapter is: what does design begin with? The answer unfolds in this order: grasp the full picture of the elements, choose where to dig, and settle the list of what will be made.

1. The Elements of Design: The Full Map

First, we lay out the full picture of the elements that make up a world. The objects of worldview design can broadly be organized into eight domains: geography and environment, history and chronology, natural law and the supernatural, society and institutions, economy and industry, culture and customs, language and nomenclature, characters and factions. Each domain further branches into detailed items. Society and institutions alone contains status, law, administration, the military, education, and religion beneath it. This map's use is not completion but inspection. No world has ever filled all eight domains, and none needs to. In keeping with the principle of inversion of priorities, a worldview carries waste, but even that waste must be managed on the map. The map's practical function is threefold. First, catching omissions. It checks whether a domain a given type requires is sitting empty. A world that claims to be wuxia while leaving the factions domain empty is a wuxia with no jianghu. Since the combination of required domains differs by type, the knowledge from Typology gets cashed in here as a checklist. Second, diagnosing imbalance. A creator's taste is bound to lean toward particular domains, so only with the map do you catch early a state where a century's worth of timeline has piled up while the economy stays blank. Imbalance itself is not a sin. Only unknown imbalance is. Third, a basis for division of labor. When several people take part in design, the domains become a natural unit for dividing the work. The map is not an order to fill everything in; it is a dashboard that maintains awareness of what has not been filled. The list of blanks is at once the list of risks and the list of opportunities.

2. The Four Axes: From Classification Criteria to Design Variables

Onto the map of eight domains, we now lay a single coordinate system. In Typology, we classified worlds along two axes: time and space. Now we complete the four axes by adding natural law and philosophy to those two. The time axis handles when the world is and how time flows in it; the space axis handles where the world is and how large it is; the natural-law axis handles what is possible and impossible; the philosophy axis handles what that world holds to be right and worthy. The axes that served as classification criteria in Typology change status here into design variables. When classifying, we read a completed world as a single point on the axis; when designing, we set the axis's value directly. Typology already covered the first three of the four in full, so here we take up only the newly arrived philosophy axis. The philosophy axis is the world's value system. What counts as honor and what counts as shame, what death is taken to mean, which is respected more, strength or goodness. This axis determines the characters' standards of judgment and the makeup of conflict, and even the same event splits into tragedy or comedy depending on the philosophy axis's value. A world where death in a duel is honor and a world where it is waste stage the same scene in opposite ways. As samurai fiction proved, a single system of norms builds an entire genre. The design implication is the explicit settling of all four axes. Early in design, write one paragraph per axis, a document declaring this world's values. These four paragraphs become the upper constitution for every detailed setting that follows. Whenever judgment splits at the level of detail, returning to the axis declarations to check against them keeps the world's grain intact.

3. Entry Point: When Does Design Begin?

With the coordinate system erected, the next matter is timing. The point at which the designer is brought into the world differs from project to project, and a different point of entry means doing different work even under the same job title. The entry point divides into three. First, entry at the planning stage. This is the case of building a world from a blank page with nothing on it, the textbook path for a deliberately-designed worldview. Freedom is at a maximum but verification is nonexistent, so a structure for getting a quick market read through a small deliverable has to be designed alongside it. Second, entry mid-serialization. This is the case of climbing aboard a work that is already alive and moving, the most common and the hardest. Settings are already piling up, and the deadline never stops. The first task here is not supporting creation but stopping the bleeding. First erect a registration gate for new settings to slow the rate at which contradictions multiply, and only then move into sorting out existing settings. Reverse the order, and the speed of the mess outruns the speed of the cleanup. It is the same principle as an emergency room: stop the bleeding first, surgery comes after. Third, entry after completion. This is the case of reorganizing a finished work into an IP, the after-the-fact systematization of a naturally-arising worldview. There is no time pressure, but in exchange it has to deal with the variable that the original creator's memory has gone cold and fanon has already expanded its territory. Cleanup after completion is therefore archival work and diplomacy with the fandom at once. The design implication is self-diagnosis of the entry point. Settling which of the three phases this is right now should be the first line of the work plan. The first deliverable differs by phase: for a blank page, the axis declarations; mid-serialization, the registration gate; after completion, a full-inventory survey plan.

4. Whether a Source Work Exists: Two Starting Points

Overlapping with the entry point is a separate matter: whether a source work exists. There are broadly two starting points for design, and the two starting points call for opposite processes. Design with a source work begins with unearthing. Whether it's a novel or a game, the already-existing work is the world's primary historical source. Rereading the work to dig out its settings, cross-checking scattered mentions against each other, and judging what's implied between the lines is the whole of the early process. This process's success or failure hinges on the discipline of respecting the source work. Whoever does the unearthing is tempted to sketch their own imagination on top of the historical source, but the moment the compiled document diverges from the source work, the compiled document's authority disappears. A document that mixes unearthing with invention becomes a document in which no entry can be trusted. Design without a source work begins with decision. There is no historical source to dig, so everything has to be decided, but everything cannot be decided at once. So the order of what gets decided first becomes the core of the process, and the answer is the four axes covered above. Declare the axes' values first, then decide the details by holding them against the axes. The two starting points also split the character of the documents. For a source-based document, source citation is its lifeblood, so every entry carries a supporting installment number and page; for a document without a source work, decision history is its lifeblood, so every entry carries a decision date and a rationale. The design implication is the boundary of the mixed type. Most real-world work sits in a state with a little source material, and here the unearthed portion and the newly decided portion have to be managed as distinct, marked apart by notation. Simply tagging source-grounded entries apart from newly confirmed entries prevents most of a document's trust problems.

5. Progressive Concretization: Where to Dig First

Once the starting point is fixed, digging begins. A worldview is not completed all at once; it concretizes itself progressively. The problem is direction. Digging every area to the same depth is impossible, so a priority for where to dig deep first is needed. The standard has already been given: the crossroads named by the principle of detail, the spot the story will pass through often. Translated into process, this becomes demand forecasting. Catalog the stages of planned works, the protagonist's affiliation and movements, and the institutions and factions that will recur, and that catalog is the list of priority excavation points. Without a demand document, priority ends up being set by taste. The method of excavation is the method of the wastefully-minted coin. Take one object standing on a crossroads and dig into its history, and the chain running from currency to mine, from mine to taxation, pulls up the surrounding settings along with it. Digging along the chain is more efficient than filling individual items in isolation, because it doubles as consistency verification. What needs watching is the upper limit on depth. A hole once started is interesting enough that digging continues, but depth with no demand behind it is where setting-play begins. The design implication is a time-box for the unit of excavation. Decide in advance the amount of work assigned to concretizing one point, and once it's spent, move to the next crossroads. A rhythm that consciously alternates between a phase of shallow widening and a phase of deep digging is the practical form progressive concretization takes. Whoever digs without rhythm is invariably found still in one hole.

6. Compound Expansion: How the Scope Widens

If depth is progressive, breadth is compound. A worldview's scope does not grow in one direction; it expands compoundly. This is because one setting summons the settings of adjacent fields. Fix the currency and economy gets summoned; fix the economy and the nation follows; fix the nation and diplomacy and war follow. Seen on the map of eight domains, it's a structure where an item in one domain sends out tentacles into the others. This expansion is both the source that gives a world organic coherence and the point where control is most easily lost. Answer every summons, and the work diverges. Setting out to fix one currency and ending up writing the trade history of an entire continent is this job's common shipwreck. The one who runs aground is usually a conscientious person. The more conscientious the person, the more they need a control tool. The tool of control is differential resolution. It's applying, to the interior, the low-resolution technique that was learned for handling boundaries. Respond to a summoned adjacent setting first at low resolution — with only a name and a one-line summary — and leave an unelaborated marker in its place. The list of markers becomes the queue of candidates for the next excavation. This way the chain gets recorded without the work diverging. The design implication is making the history of expansion visible. Tally which direction the world grew in and by how many entries this week, and the diagnosis of imbalance and the next priority call get made on data. Expansion is not something to block; it's something to put on the dashboard. The homework Chapter 1 assigned — managing waste — gets a concrete tool here.

7. The Difference Between Research and Worldview Work: A Test

The process description up to this point lets us settle an old confusion: the distinction between research and worldview work. The principle was settled in Chapter 1. If exactly what's needed and no more is what's asked for, it is research; if it carries waste as a required element, it is worldview work. Translated into a practical test, this comes down to looking at the nature of the request. Is the request closed, or is it open? A closed request has an answer whose shape is already fixed. A request to compile the organizational structure of the late-Joseon Podocheong — the era's police bureau — has a correct answer, and the work is completed by collecting that answer. An open request requires the answer to be generated. A request to build this world's system of law and order has no correct answer; decision-making and managing that decision's ripple effects become the body of the work. Even within the same project, work of both natures comes mixed together. Worldview work for historical fiction is mostly closed requests, work for fantasy is mostly open requests, and faction splits exactly down the middle. The inside of a gap map corresponds to open requests; the fixed points outside it correspond to closed requests. The reason this distinction matters is that process, staffing, and estimates all differ. A closed request is closer to a researcher's job, since research capability and source management are the core of it; an open request is the designer's job, since decision-making and consistency management are the core of it. The design implication is classification at the intake stage. Simply tagging every incoming request as closed or open makes the allocation and forecasting of work considerably more accurate. The confusion of handing closed work to the designer and open work to the researcher is a major cause of attrition in this profession.

8. The Spec List: A List of What Will Be Made

The document that ties the process's control tools together into one is the spec list. A spec list is the list of deliverables confirmed for production in this project. The key point is that it is not the whole of the world, but the list of what will be made this time. Entries are written by deliverable unit: one continental map, an overview document for five major factions, detailed setting for the protagonist's affiliated organization, a draft timeline, a first pass at the glossary. Each entry carries a resolution grade, a deadline, and an owner. This document's effect runs in three directions. First, it is a contracting device with the client. A request to "build a worldview" has infinite scope and is therefore a seed of dispute, but once the spec list is signed off, the scope becomes finite. Additional requests get handled as a revision to the list, leaving a basis for the estimate. Turning infinite work into a finite contract is this document's first reason for existing. Second, it is a device for managing waste. Work outside the list is not forbidden. But waste done knowing it is outside the list is different from waste done without knowing, and only the former is managed waste. Third, it is a dashboard of progress. The impression that worldview work is a job with no end in sight erodes morale, and the list artificially manufactures the state called completion. You need a list of things that end in order to endure the thing that never ends. The design implication is periodic renewal of the list. A spec list is not a document written once at kickoff; it is a living contract revised again and again to reflect the history of expansion.

9. Deliverable Specifications: The Conditions for a Deliverable Form

For each entry on the spec list to be recognized as complete, it has to pass specification. If deliverables-first was the declaration, here we settle the content of that specification. The conditions for a deliverable that can actually be handed off were: searchable, citable, traceable. Translated into a document specification, this becomes the following. First, itemization. A deliverable must be a set of entries with headings, not a block of running prose. Search only works if there are entries. Second, regularity of headings. The same subject is always written under the same name; if the notation wavers, search may exist but trust does not. Third, source or decision history. If it's based on a source work, the supporting location must be attached; if it's a new decision, the decision date and rationale must be attached to every entry — only then does citation work. Fourth, version notation. The whole document and each individual entry need a version for traceability to work, so the recipient can confirm the freshness of their own copy for themselves. Fifth, the presence of a summary layer. Not every reader reads the full text, so every entry needs a one-line summary and every document needs an overview. These five conditions do not require an elaborate document. A single spreadsheet can satisfy all five, and in practice the standard tools of the trade are the wiki and the spreadsheet. Specification comes not from the flashiness of the tool but from satisfying the conditions. The design implication is settling the specification early. Agree on this specification before the first deliverable ships, and every piece of work that follows becomes automatically compatible. Applying a specification retroactively costs several times what applying it from the start does.

10. The Work Cycle: Extraction, Definition, Registration

The map, the coordinates, the list, and the specification are all in place. Last, we settle the basic circuit of work that runs on top of them. The minimum cycle of worldview work is a rotation of three actions: extraction, definition, registration. Extraction is the act of pulling a setting candidate out of a source. From a sentence in the source work, from a pile of research material, from something said in a meeting, a fragment with the potential to be incorporated into the world is fished out and put on the list. Definition is the act of settling the meaning of the candidate that was fished out. It decides exactly what that thing is and is not in this world, and what name it goes by. Registration is the act of putting a defined entry into the master canon document. Only an entry that passes a conflict check against existing entries gets registered, and only what is registered becomes an official member of the world. This is exactly the practical definition of canon: the set of entries that have passed registration. This cycle repeats regardless of scale. In the first week of designing from a blank page and in the maintenance of a ten-year-old world, the circuit that turns is the same. We can confirm that every phase seen earlier was a variation on this circuit. Unearthing a source work is a cycle where extraction is oversized; sourceless design is a cycle where definition is oversized; and stopping the bleeding mid-serialization was the emergency installation of a registration gate. Each of the three actions has its own system of technique, so the three chapters that follow take one apiece. The technique of extraction is the keyword method; the technique of definition is definition and naming; the technique of registration is the data pipeline. One circuit, split and packed into three toolboxes.