Chapter 4. The Interface between Vault and Writing — One Source of Truth and Bidirectional Synchronization

Before writing, always open the Vault. This is principle, not habit. Relying on memory lets the K-pop K inside one manuscript drift from the official world until manuscript 50 contradicts manuscript 1.

After writing, always update the Vault. A draft discovers how a person speaks under pressure, which procedure the world uses, or how a place smells. If such information stays only in that manuscript, later work cannot use it.

These two directions form the interface between Vault and writing.

SSOT: Single Source of Truth

Software development uses the SSOT principle: one official version of any datum. Duplicate official copies eventually disagree.

When Marvel made Wakanda Forever, Wakanda's currency conflicted with Black Panther (2018), requiring continuity work; Marvel subsequently centralized canon data. During Forgotten Realms' transition from fourth to fifth edition and its fifty-year timeline jump, Wizards of the Coast designers likewise used a central wiki as reference for new novels and rulebooks.

In MEJE's hundred-piece project, the Vault is SSOT for world rules, member traits, events, and spaces. A manuscript is not.

Manuscripts consume the Vault and supply new data back. The correct direction is not “the manuscript says this, therefore the world is so,” but “the Vault defines this, therefore the manuscript follows it.” The world constrains writing.

The exception is discovery. If writing a member reveals an occupational custom not yet defined, this is not deviation. Add the discovery to the Vault, where it joins the SSOT.

What to Take from the Vault

Before writing, check six library headings introduced in Librarying: background, material, people, events, objects, and places.

• Background: time and space—practice-room size, light color temperature, floor material, time of day. Without this, fifty manuscripts invent fifty different rooms.
• Material: whether the subject appeared before. Repetition is not inherently wrong, but it should be chosen knowingly.
• People: relationships, speech habits, past events, and physical expression, preventing characters from drifting.
• Events: timeline conflicts and contradictions with other manuscripts.
• Objects: official definitions of symbols, goods, and tools.
• Places: descriptions consistent with prior appearances.

A short sketch may check them lightly, but repairing one skipped contradiction costs more than checking all six first.

From Manuscript to Vault — The Feedback Loop

After writing, either the manuscript merely expressed existing Vault information—no update needed—or it created new information, which must feed back.

Four common discoveries are:

1. New world detail: a corridor bulletin board and its notes, added to the place entry.
2. Character discovery: Myeongsuk bites her left thumbnail under extreme tension, added as a telling detail.
3. Relationship discovery: dialogue reveals shared history absent from the Vault; check against other manuscripts, then add it to relationship records.
4. World-rule discovery: for example, members under a public-activity suspension must follow a lodging dress code when going out; add it to world rules.

The feedback loop makes each manuscript enrich the Vault, which then feeds the next manuscript. Stage 2 of MEJE's seven-stage pipeline builds the Vault; Stage 7 consumes and refills it. As long as the loop continues, the IP grows.

Writing Confidentiality Principle

Internal filenames, version numbers, procedure names, and Librarying system code names must never appear in public manuscripts.

A story is a world product, not a production report. A line such as “This manuscript satisfies Criterion F in procedure v4.2” fails immediately. Readers neither know nor need to know the workflow.

The principle protects two things. First, reader experience: production machinery blocks the window into the world. Second, intellectual property: MEJE's internal methodology should not leak unconsciously through outputs. Chapter 14 revisits this in AI collaboration.

Parallel Writing and Vault Conflicts

With one hundred pieces, several writers—or human and AI—may draft simultaneously. The largest risk is that one piece treats an event as future while another treats it as past.

Prevent this by sharing the Vault in real time. Any writer who adds a world detail records it immediately and informs the team.

If real-time sharing is unavailable, establish a Vault snapshot before parallel work. Everyone writes against it, then conflicts are reviewed during integration.

This resembles version control. The Vault is the master branch, each manuscript a feature branch, and the feedback loop after completion is the merge where conflicts are resolved.

The Writing Interface in Practice

Suppose we write a comic sketch about a fictional K-pop K member named Haneul.

1. Open the Vault and spend five minutes reading Haneul's name, age, position, personality, speech, past events, and relationships.
2. Choose one everyday cross-section from event and material records; check duplication in the Coverage Matrix.
3. Record stakes, viewpoint, form, and length on the design sheet as Chapter 3 directs.
4. Write with the design sheet open; consult the Vault whenever detail is needed instead of relying on memory.
5. Run the six quality criteria and obtain a pass.
6. Add discoveries to the Vault and update the Coverage Matrix.

One hundred pieces repeat this cycle one hundred times.

When the Vault Is Weak

Some of the hundred characters may have sparse records. There are two options: strengthen the Vault first through Librarying, or let writing build the Vault as an exception.

Under the second option, every invented detail must enter the Vault immediately. Once added, the Vault becomes the standard for all later use, preserving SSOT even while the manuscript leads.

Start weakly documented characters with a sketch or four-panel scenario. Short forms make fewer discoveries and reduce feedback burden. Move to short stories as understanding grows.

What Chapter 4 Conveys

The interface is methodology, not intuition. Its purpose is not to mechanize creativity, but to make one hundred manuscripts share one world.

Each piece must be good alone and become larger beside the other ninety-nine. The Vault protects and grows that world: writing consumes it and enriches it.

The next chapter turns to the hundred protagonists—how to manage that number, choose each person, and operate the Coverage Matrix.

---

- Main Text / Library-Based Storytelling 100, Session 4 / 2026-04-30