Files
magic-factorio/tasks/tech-tree-plan-v0.2.md

5.5 KiB
Raw Permalink Blame History

Implementation Plan: Magical-Law Progression v0.2

Overview

Implement the progression in docs/tech-tree-v0.2.md as a sequence of playable magical-law experiments. The highest-risk mechanic—the portal's bounded random matter exchange—is prototyped before large content production. Once that proves fun, each ecology zone adds one distinct simulation law and one automated proof loop, culminating in a player-built dimensional gate and post-victory entropy economy.

Architecture decisions

  • Existing belts, mana pipes, bridge endpoints, fixed ticks, and event wakeups remain the physical substrate.
  • A biome law changes simulation through explicit data/state contracts; it cannot reach into arbitrary scene nodes.
  • Probability uses deterministic seeded result bags over mass batches. Saves preserve the active bag and seed so reload cannot reroll outputs.
  • Failure states are recoverable resources or paused machines on standard difficulty, never silent deletion.
  • Research proofs are produced by satisfying observable state sequences, not by renaming ordinary assembly recipes.
  • Boundary crossings are explicit logistics nodes. The visible map stays continuous while the transport graph remains partitioned.
  • First warp is deterministic once its conditions are met. Random exchange begins immediately after victory.

Dependency graph

M1 portal fun prototype ------+
                              |
M2 biome-law contract --------+----> M3 catalog + validator ----> M4 proof runtime/UI/save
                                                                  |
                                                                  v
M5 runic opening -> M6 life/death -> M7 ember -> M8 mirror -> M9 first warp
                                                                    |
                                                                    v
                                                         M10 postgame portal loop
                                                                    |
                                                                    v
                                                           M11 pacing proof

Task list

Phase 1: Prove the risky fun

  • M1: Prototype bounded random exchange, sorting, entropy, and recovery
  • M2: Define biome-law and boundary-crossing contracts
  • M3: Load and validate the 80-node v0.2 catalog
  • M4: Implement experiment-driven research proofs, UI, and saving

Checkpoint: Core hypothesis

  • A ten-minute portal toy creates at least two meaningful strategies: stable targeting and high-entropy throughput
  • Reloading cannot reroll a result bag
  • Each planned ecology law has an isolated state model and a recoverable failure path
  • Mainline graph is acyclic and contains all five magical verbs

Phase 2: Playable ecology progression

  • M5: Deliver the old-kingdom runic opening
  • M6: Deliver forest and undead zones in either order
  • M7: Deliver the life/death merge and ember-state loop
  • M8: Deliver mirror folding, visible probability, and identity anchoring

Checkpoint: Magic identity

  • A stripped-art test can distinguish all four external zones by rules alone
  • A starting-zone blueprint cannot run unchanged in another zone
  • Every proof line can be automated and exposes a precise stop reason
  • Life and death routes remain order-independent and converge without duplicate blockers

Phase 3: Victory and continuation

  • M9: Deliver the twelve-segment gate, synchronized ritual, and first-warp achievements
  • M10: Turn the opened gate into the postgame matter-exchange and paradox-research loop

Checkpoint: Complete loop

  • Completing G09 alone does not win; the walker construct must actually cross the aperture
  • First warp has no random failure
  • Continuing after victory immediately exposes useful portal production
  • Portal exchange reduces one resource pressure while measurably increasing sorting, buffering, entropy, or mana work

Phase 4: Timing proof

  • M11: Instrument and tune the expert and normal routes

Checkpoint: Release candidate

  • Expert-route P50 ≤ 7:15 and P90 ≤ 8:00
  • Experienced first-play P50 is 1821 hours and P75 ≤ 24:00
  • The expert route still demonstrates growth, decay, three-state conversion, echo separation, and four-anchor synchronization

Risks and mitigations

Risk Impact Mitigation
Random exchange feels punitive or unplannable Critical Prototype first; use bounded bags, visible ranges, bias controls, and recyclable outcomes
Ecology laws become four unrelated minigames High Keep common item tags, mana frequencies, proof records, and boundary-node interfaces
Magical layouts become tedious pattern copying High Ship fixed 3×3/5×5 templates with snapping and previews
Portal becomes strictly better than gathering High Cap early usable mass at 90% and exact targeting at 65%; charge entropy and catalysts
Eight-hour play skips the interesting systems Critical Validate the mandatory closure and time the five explicit mechanic demonstrations
Region effects cause per-cell/per-frame cost High Recompute zone membership on construction and wake entities only when law state changes

Explicitly deferred

  • Full combat and tower-defense balance
  • Procedurally generated infinite biome types
  • Multiple full-sized playable dimensions
  • Freehand rune drawing
  • Irreversible corruption or permanent random destruction on standard difficulty