Hold the Flood — Design Review

REV103 · r5 · 2026-09-06 16:40 UTC · Independent Design Critic · In review — PMU004 applied to this review: the owner has redirected the scope from one district to a campaign of many, so REV103's remaining items become integration issues rather than gates, and the criteria this role will apply to the next owner-facing milestone, DES104, are set out now so the campaign is written against them. Not a review of a new document revision; the Design Document and World and Mystery are unchanged since REV103 r4. REV103 r4–r1 are on the REV103 archive; REV102 r6–r1 on the REV102 archive; REV101 r8–r1 on the REV101 archive; every revision is in page history (REV103 r4 711ef4 · r3 dcd90e · r2 3ff337 · r1 285925 · REV102 r6 538724 · … · REV101 r1 7ad187).

What was checked

What PMU004 changes for this review

1. The scope, not the criteria. Criteria 1–11 stand; the owner has changed what the complete product is. Every verdict this page gave on "the owner packet" concerned the one-district pitch, which the Expansion Brief now calls obsolete as a product; those verdicts are moot, and no further closure pass on that pitch will be issued unless PM asks. Nothing in the rule-level findings is moot: the spine rule, the precedence, the cables, the rack, the charge ledgers and the mystery chain are exactly what the mechanics page calls "the shared rule contract" that districts must not secretly change.
2. REV103-2, -7 (the weir clause), -8 (one sentence) and -9 (what the relay's channel carries) remain open as integration items, as the Expansion Brief says of PMR003/REV103 contradictions generally. They will be checked when the lead and the world designer revise, on the exact revision, and not chased between times.
3. REV103-2 is now confirmed twice. The External Critic's X27 reproduces the 9 / 9 tally independently. The lead has not yet read either.
4. This role's assignment is unchanged. Design Index: critique exact linked submissions; the four proposed designers' prompts are not yet posted and their roles are not assumed to exist. PM-seeded snapshot pages are reading aids, not submissions, until an author adopts them; they will not be critiqued as design.

REV104 — the criteria this role will apply to DES104, stated now

PM asks that the campaign package be reviewed "for appeal and coherence". Appeal is the owner's; this role can review whether the appeal claimed is supported by designed choices. The check, so the campaign can be written against it:

The shared rule contract, first. One statement, before any district: conduction, power, attraction and charge as they now stand, the retarget-or-spread precedence, what counts as a beacon, the two cables, the rack rules, the once-per-organism yield. Every district exception must name its readable signal and the choice it improves (the mechanics page's own rule). A district that changes what power or signal means fails here regardless of its other merits.

Per district, the Expansion Brief's six, read as this role reads them:

• a distinct route or machinery problem under the shared rules — the test is a written expedition in that district whose choices could not occur in the pumping district; "eight renamed pumping yards" is PM's phrase and the right one;
• a reason to prepare a different loadout, with a viable alternative and a real tradeoff — the test is two drawn racks for the district that differ in what they do, not in a number; the six-slot cap and seven modules must be justified or revised for this, as the brief requires, and more items must make new choices;
• enemy combinations that change movement or priorities — the test is a precedence question the combination poses that a single enemy type cannot; health and tint do not count;
• a local discovery and a contribution to the larger mystery whose evidence changes a later action — the false-refuge chain is the model: observable clues, a fair inference, a free experiment, a consequence, a recovery path if missed, an answer for replays;
• a consequence for another district, the refuge or a return — stated as a rule, not a promise;
• asset categories and gaps — no availability claimed unless linked and checked, as before.

The campaign as a whole: a district map with reasons to travel and to return; early, middle and late equipment and combat choices across the campaign, not a damage catalogue; a colony loop that changes preparation and campaign choices without the upkeep tax the baseline forbids; a mystery thread across districts with the same fairness the local one has; the pumping district's finale reframed as a local outcome, as the brief suggests, with its three endings' district-wide consequences restated for a campaign; a content coverage table and asset burden; hypotheses labelled; no hours figure produced by multiplying the old district's estimate, and first discovery, required revisits, optional exploration and replay separated.

Method: per-beat ledgers for every written expedition, as for REV103; and for every district a short "shared-rule diff" — what it adds, what it changes, what signal announces the change.

Deliverable, per PM's own wording for REV102: the five most consequential remaining issues, or an evidenced readiness verdict, distinguishing consistency and specificity from enjoyment, balance and feasibility, which no document can establish.

What carries from the one-district work, so it is not lost in the move

The spine rule and its precedence, verified across nine revisions; the visible cables; the rack and its leak rule, with the ledgers that now close to the unit; the false-refuge chain, which is the campaign's template for a fair local mystery; the finale's machinery, which becomes a district outcome; the honesty apparatus (§16, decisions with costs, observed versus predicted). The owner found the premise promising on the strength of these. The campaign should be built by combining them, not by relaxing them.

For PM — three pointers, one line each

• The Expansion Brief cites Team / Expansion Prompts as ready; it is 404. The Design Index lists Cutter, Equipment, Enemies and Colony as populated; they are 404. (The reader's "do not stamp a pointer with something that will move" convention, in reverse: do not point at something that has not arrived.)
• Hub and Tasks still state the one-district scope and the old DES103/REV103 framing; until they are updated, PM's canonical pages disagree with PM's brief.
• No owner-approved district count exists; eight is PM's drafting target. This review will treat any hours or district figure as a labelled hypothesis until the content map exists.

Finding index — where every finding lives

Superseded reviews: REV103 r4 · r3 · r2 · r1 · REV102 r6 and earlier on the REV102 archive · REV101 r8 and earlier on the REV101 archive.

PM report — REV103 r5

• Task: REV103 (its residuals now integration items under PMU004); REV104 criteria stated. Artifact: Design Review, REV103 r5 (this page); r4 moved to the REV103 archive verbatim.
• Checked: Expansion Brief 1a568e, Design Index 3cc078, Mechanics / Power and Attraction 782139 in full; the cited pages' existence; External Critique a16667's X27 against REV103-2. Design Document and World unchanged.
• Changed decisions: none by this role.
• Evidence: the pages quoted; the 404s; nothing built, prototyped, tested or created; no external source verified or repeated as fact.
• Remaining risk: that the campaign is drafted by relaxing the verified rules rather than combining them; the shared-rule-diff check above is the mitigation.
• Next handoff: lead and world designer — the four integration items on their next revisions; PM — the three pointers, and the Hub/Tasks update. I review DES104 when it exists and post nothing before then except a PM correction or a revision that closes the integration items. One exact PM request: confirm that DES104 will be reviewed against criteria 1–11 plus the Expansion Brief's per-district six, as set out above, so the lead and the campaign designer write to one list.