Hold The Horde — Exploration Structure

Status. A provisional proposal for review, not owner-approved. It describes how repeated expeditions can form one discoverable, revisitable station. Example fiction is labelled where used. The revival explanation, the station's full history, the escape solution and the languages' design are not settled here and are not needed to settle this. No layout, district count or content programme is proposed.

The shape of the campaign: reach, not unlocking

Oxygen limits how far the astronaut can go and return (owner requirement). That makes reach the natural structure for a campaign of many expeditions: at any moment there is a band of the station the player can work in, and a frontier beyond it.

The frontier moves when the cost of getting somewhere falls. Three things lower it: carrying capability, so a canister goes further; knowledge, since a known direct route is cheaper than a searched one; and infrastructure the player restores, such as power to a hoist.

Barriers and reach are not alternatives. A sealed door is shut, and opening it changes access; restoring a hoist changes access and travel cost at once. The emphasis is that barriers are station state the player can act on with what they understand, rather than locks waiting on a key found elsewhere. Nor is the reachable set a ring drawn around the ship: oxygen limits time and capability, so reach is shaped by route, hazard and load, and can run deep along a cheap path and shallow across a costly one.

Reach also depends on spare canisters competing with equipment and salvage for capacity — a rule proposed in Oxygen and Observation and Build Comparison, not owner canon. This structure does not depend on it holding.

This has a consequence worth stating plainly: explored space becomes the corridor to unexplored space. The player will cross early areas many times, so those areas must not be inert.

What is authored and what varies

Deliberate infrastructure work by the player — power restored, a section unsealed — persists, including its side effects: once a section is no longer isolated, it stays that way. Consumable state, such as a cleared room, does not persist.

Holding geometry fixed is a preference, not a brief requirement: the brief's list of what a revisit may offer is not exhaustive, and a changed environment could reasonably include a changed layout. The reasons for choosing it are given under the alternative below.

Where an expedition's goal comes from

Every sortie leaves with something the player could say in a sentence and check on arrival. The goal is orientation, not a quest list, and it can be abandoned or replaced mid-trip: "clear immediate goals" asks that the player always knows what they set out to do, not that they must finish it.

Goals come from five ordinary sources — something recorded and not understood, a cache left behind, a person not yet sent home, a capability the colony wants, a named place not yet visited. Exploration itself is a legitimate goal provided it is specific enough to be finished: "go further" is not one, but "reach the far end of the painted names and record whatever marks the wall past them" is. Completion evidence stays modest: an object in the pack, a facsimile in the journal, a person aboard, a mechanism running. None of it needs a quest system.

Finding routes and recognising places

Places are recognised by what is in them: a machine of a particular kind, a wall of marks, a room where someone lived. Recognition is authored, not incidental.

The station also carries its own addressing. (Proposed fiction.) The human crew painted section markings for their own use, in their own hands, so places have readable in-world names the player can write down and say to a survivor. This makes location a thing the player learns rather than a coordinate the game hands over, which is what makes personal note-taking worth doing.

That fiction has a limit, and the limit is a design dependency rather than a flourish: painted names cover only where the crew wrote them. The journal's location reference must therefore rest on a reliable substrate, with the painted names as a legible layer above it — otherwise references break exactly where exploration becomes most interesting.

What that substrate is, and what the suit always shows, belongs to a shared navigation and inspection contract owned by Systems and not yet published. This article assumes only that a recorded location stays resolvable without a painted name. Always-readable survival state, an unknown route, a recorded location and a learned crew name are four different things; intentional uncertainty about where a route leads is not the same as hiding survival information.

How earlier observations gain new meaning

The established pattern is that understanding reinterprets marks the player already walked past. Applied across expeditions, the same evidence keeps addressing a larger object.

First a cut mark describes a component. Then a room: a bay whose contacts all carry one family's cut mark was built to do one job. Then the station — but here the evidence changes hands. A run of doors carrying painted copies is not the builders' scheme; it is a boundary a person drew, using a borrowed shape, and reading it tells the player about that person rather than about the machinery. Keeping the two kinds of mark distinct is what makes the third tier a discovery about people. The player ends up reading the layout itself with evidence gathered for a different purpose, and no new system is required for it — only that the marks stay where they are.

Three expeditions

Labelled example fiction, using the accepted sortie as its first step. Observed clues, available inferences and resulting choices are kept separate; the inferences are what a player could conclude, not what the game states.

First — the errand. Observed: a rack shutdown runs through one cut-marked part, drains at a second, stops at a third; a pressure door bears an irregular hand-painted copy of one of those shapes; aboard ship, Yeva recognises the paint as Dorn's. Available inference: the marks name families rather than parts, and a person deliberately sealed that door. Choice: treat the door as a place to return to, or bank the trip and go elsewhere.

Second — the sealed section. Goal: reach the door and find out what it closes. Observed: the same hand has painted the same shape on several fittings along one route, not on the door alone; behind an open panel, the section's power feed is dead because a contact has been physically pulled from its seat, and the seat itself carries a cut isolator mark. Available inference: the sealing was deliberate and done by hand rather than by failure. Whether the painted fittings were marked in an order, and in which direction, is not observable from the marks themselves — that reading stays tentative unless the fiction gives it a physical basis, such as later paint overlapping earlier where two marks meet. Someone working outward and closing behind himself, and someone shutting something in, both survive the evidence; the design does not decide between them. Choice: re-seat the contact and open the section, accepting that whatever the isolation was for is no longer isolated; or leave it and follow the painted fittings the other way.

Consequence, route: re-seating the contact is a hand action at the panel — the same work the mark's original meaning describes, not a key or a found item. The restored feed also powers a hoist on the main run, and both effects persist: the section stays un-isolated and the hoist stays powered for every later expedition, shortening the trip inward. Capability gained by understanding what a mark meant.

Third — the last painted name. Goal: reach the far end of the painted section names and record whatever marks the wall past them — checkable on arrival, and abandonable if something better appears. Observed: past the hoist, painted names continue in a different hand, then stop. Beyond them the cut marks carry on across space carrying no paint at all. Available inference: the crew's labelling stops here. Whether they never went further, went and stopped writing, or went and did not come back is not settled by absent paint, and the article does not claim it is. Choice: press on without painted names to navigate and record by, or work the named band thoroughly first.

Consequence, narrative: the crew's own record has an edge, which is a discovery about people rather than about a mechanism, and it sharpens a question instead of answering one. It resolves nothing about the station's origin and is not intended to.

What a return visit changes

A revisit is not a repeat because the population and state differ, because infrastructure the player restored is still restored, and above all because the player reads more than they could before. The strongest version costs no new geometry: the room is identical and the player is not.

Journal references and cache recovery

Fixed geometry is what makes both work. A recorded location stays valid indefinitely, so comparing a mark against one seen twenty expeditions ago is reliable. A cache sits where it fell and the route to it continues to exist; whether it can currently be reached, carried and returned within supply is a question of reach, not of whether the room still exists. Those recovery rules are proposed in Death and Persistence and remain under review; nothing here depends on their final form beyond the site persisting.

The alternative, and why it might be better

Fixed anchors, variable connective tissue. Authored destinations keep their contents and marks; the corridor and service space between them is regenerated or re-dressed between expeditions.

Its case is stronger than a first reading suggests, and it does not lose priority one. Because anchors keep their contents and marks, a sealed door that is an anchor is exactly as reinterpretable as it is here: the player can still walk past it, learn what its mark means, and come back to a door that is still there. Location references to anchors keep working, and the crew's painted names can live on anchors too, so nothing forces the game to hand over coordinates. What varies is the tissue between anchors, and the loss is confined to that tissue: an incidental crate two junctions along may not be there next time, and route memory stops accumulating.

It is also plausible — though not established — that this saves authored content and keeps traversal varied by construction, without leaning on population and state to supply all the novelty.

Why fixed geometry is still preferred, as a judgement rather than a deduction. Under anchors, the designer must nominate in advance what is allowed to carry meaning. Under fixed geometry, anything can: a mark on an incidental crate, the arrangement of a room nobody flagged, a route that turns out to be a boundary. For a game whose first priority is noticing what nobody pointed at, that difference is worth paying for. It is a preference about where discovery is richest, not something the brief settles.

The costs accepted. Authored content burden, which remains the most credible threat to the brief's scale ambitions. The risk that a permanent station becomes a solved commute once routes are known — and because restored infrastructure persists and shortens known runs, difficulty likely has to come from depth rather than from re-crossing settled ground. Varying population, hazard and power state is the proposed answer to staleness. Each of these is a prediction; none is measured, including the comparative content cost of either arrangement.

What would change this preference

Evidence that the authored burden for a permanent station is infeasible at the intended scale, or that varying state and population does not keep re-traversal interesting. Either would make the anchor arrangement the more honest one. No such evidence exists, and written description cannot produce it.

Unresolved

Map generation and how much of the station exists at start; enemy and mechanism reset rules beyond the table above; the journal's positional substrate; how much reach a capability adds; the revival cause, station history, escape solution and language design, all untouched. Nothing here establishes fun, pacing, content feasibility or the 30 / 90 / 100–200 hour ambitions.