Big Computer Games (1984, David H. Ahl / Creative Computing Press)

Author(s): Edited by David H. Ahl. Individual games by named contributors (Kenneth R. Murray, Richard R. Galbraith, Lloyd Johnson, Stewart F. Rush, and others). Art direction by Patrick Calkins. Sources: The book itself (Creative Computing Press, Morris Plains NJ, copyright 1984, ISBN 0-916688-40-2, LCCN 84-12730), read from a scan of the print edition. The scan carries no text layer, so pages were read as high-resolution renders and the essays additionally via OCR; page citations are to the printed page numbers. Category: inspiration, Batch 12 (BASIC-listing corpus mining). Five games plus both design essays read in full. Cross-cutting cart mapping lives in ../synthesis-basic-games.md. Serves: runtime / engineering (assigned by Josh; see cart map)

What it is

Big Computer Games is the volume where the BASIC anthology grows up. Twelve programs, most of them long, aimed at players who want a session rather than a diversion: a nine-month election campaign, an interstellar trading career, a multi-year municipal budget, a coast-to-coast haul, a survival puzzle on the moon. The table of contents advertises them as management and simulation rather than as games.

Its most valuable content for KN-86 is not a game at all. Two prose essays, How To Write An Adventure and Tips for Playing Adventure Games, bracket the interactive-fiction problem from both sides: the first is a concrete data-model recipe for building one, the second is a player-behavior catalog that reads, inverted, as a fairness specification.

Read for this capture: Presidential Campaign, Star Merchant, Streets of the City, Survival, Trucker, plus both essays.

Mechanics

Read from the printed prose, sample runs, and listings. The longer listings were skimmed for their core model rather than traced line by line; page citations are to the printed pages.

The two design essays

• How To Write An Adventure (pp.100-102) is a build recipe, and nearly every step is a data-model decision:

  • Start from five plot questions in order: where does it take place, what is the main purpose, what kind of world, what obstacles must the player overcome, and how does the player get past them.
  • Draw the map first; rooms are a graph. Roughly 40 room nodes, each with up to six directed exits (N/S/E/W/U/D), where 0 means no passage that way.
  • Room prose is decoupled from room geometry: the description is data keyed to the node, the links are integers.
  • Objects are (object, home-room, value) triples, and home-room 0 means “hidden, nowhere yet.” A trap door sits in room 0 until MOVE RUG sets its room to a real number. That one field is the entire lock-and-key primitive: a verb mutates an object’s location to reveal it.
  • The parser splits input into a verb index and a noun index against a vocabulary table; the main loop branches on the verb index.
  • One verb routine per verb, and each accepts only a few sensible nouns; anything else gets a terse refusal.
  • Effects are state mutations plus a confirmation line: eating the food sets its room to 0 and prints a one-line acknowledgement.
  • Puzzles are dependency chains. The worked example is a lock-and-key tree: to reach the treasure you need the ring, to get the ring you eat the pill then lift a gate, to eat the pill you open the envelope in the pantry.
  • Magic words and codes work as teleport shortcuts and gates, reachable with no other path.
  • A per-room display routine (description, contents, and all non-zero exits) is the render contract.

• Tips for Playing Adventure Games (pp.116-118) catalogs what players actually do, which inverts cleanly into design duties: they walk away and resume cold, and they trade hints socially; they read words in “quotes” or *stars* as deliberate clues and try decoding them; they accept that some obstacles exist only to confuse; they recognize that a blocker usually means a resource they have not found yet; they try seemingly stupid commands because the dumb verb is often the answer; and they save every fifteen minutes and always before anything with unknown results. The essay names the canonical pacing failure directly: a player will spend more time working around the missing hammer than looking for it.

The games

• Trucker (p.87, by Richard R. Galbraith, Creative Computing March 1981): an 18-wheeler from Los Angeles to New York against a $1,955/month loan. Setup commits you to one of three cargoes (mail pays least but has a hard deadline; oranges pay most but burn 7 gal/day of diesel for refrigeration even while you sleep) and one of three routes with distinct distance and hazard profiles (northern is shortest but has mountains, snow, and weigh stations; southern is safest and longest). Then a per-segment loop where speed is a single dial wired to four consequences at once: fuel economy (which falls off sharply above 55 mph), ticket probability, accident probability (amplified by weather and fatigue), and arrival time against the clock. Fuel is deliberately imprecise: a 5-gallon gauge granularity and a speedometer accurate to 3 mph, so refuelling is planned under uncertainty. Fatigue is a second clock racing the delivery clock: sleeping costs hours, driving tired spikes crash odds.

• Star Merchant (p.57, by Lloyd Johnson, Creative Computing Aug 1981): lease a starship (20 years, 2,000,000 credits, 20 cargo holds) and run trade routes until you retire rich, go bankrupt, or the lease expires. Every commodity carries one of five trade classes (raw, refined, industrial, agricultural, non-agricultural); each star produces some classes and lacks others, and the price routine sets a local price by whether the current star is a source or a sink for that class. Stars carry x/y coordinates, so distance, direction, and fuel cost are geometric. Three constraints squeeze simultaneously: capital, 20 cargo partitions, and the lease clock, since every jump burns time while salary and lease accrue. Notably the pricing is static per visit: your own trade does not move the local price.

• Presidential Campaign (p.47): nine monthly turns from February to October. Each month you choose whether to fund-raise or campaign, and campaigning means spending from a finite war chest in specific states, each carrying a poll percentage and an electoral weight. Six issues are ranked at setup and weight how events land. A random monthly event branches with consequences: a gasoline shortage to explain, an offer of illegal contributions that pays cash against a risk of ending the campaign in a federal penitentiary, a televised debate to accept or duck, a union endorsement, foreign criticism. Resolution is winner-take-all per state, summed to a national electoral tally, so the strategic problem is where to spend rather than how much.

• Streets of the City (p.66, by Kenneth R. Murray, Creative Computing April 1983): you are the Transportation Director of River City for a multi-year term. This is a civic budget simulation, not a driving game. Each fiscal year you set transit routes, hours, days, and fares; sell bonds in specific years subject to a debt ceiling and voter approval; set a property-tax millage; split the street fund between operations and construction; choose miles of street reconstruction and interstate build; fund maintenance and safety; buy or sell buses; and answer a union wage demand. The state is a bank of coupled indices (street condition, traffic safety, ridership, fleet age, downtown vitality, service delay), each carrying an inflation and decay term so standing still loses ground. You win by meeting a goal slate by term end.

• Survival (p.79, by Stewart F. Rush, Creative Computing Jan 1982): crash-landed on the moon with 180 minutes of oxygen and 230 units of power. The map is a move matrix M: one vector per location where M(P,1..6) gives the destination for N/S/E/W/U/D, 0 means no exit, 99 terminates, and M(P,7)/M(P,8) point at the first and last print-line of that room’s description, so geometry and prose are separate. Objects live in an O vector holding either a location or 99 meaning “carried”, with a carry cap of four. Flags gate hazards and state (meteor shower, shed open, illuminator lit, bomb deactivated), and one room is total darkness that requires the illuminator. Every move costs 5 minutes of oxygen and 5 units of power, but the drain applies only outdoors: the listed locations that require oxygen and power are the surface ones, and reaching the station changes which pool depletes. The author’s headline design choice, stated in the conclusion, is that the program contains no random events at all, making it a deterministic optimization puzzle you replay to shave minutes.

Mineable patterns

• The interactive-fiction data model, complete. Rooms as a directed graph with prose decoupled from geometry; objects as location-tagged records where a sentinel location means “not yet revealed”; a verb-noun parser resolving to indices; one routine per verb; effects as state mutation plus acknowledgement; puzzles as dependency chains. This is a whole architecture in three pages.
• A sentinel field as the lock-and-key primitive. Reveal is just “set the object’s location to a real room.” No separate discovery subsystem, no new screens.
• Player behavior inverted into fairness duties. Because players resume cold, brute-force verbs, read typographic emphasis as clues, and save before risk, a design owes them resumption, verb discoverability, a consistent clue-markup convention, and a cheap checkpoint before anything unknown.
• One dial wired to several consequences. Trucker’s speed decision is the cleanest example in the whole three-book corpus: a single number that simultaneously sets fuel burn, ticket odds, crash odds, and arrival time.
• Deliberately imprecise instruments. A 5-gallon fuel gauge makes planning a judgment rather than an arithmetic exercise. Uncertainty as a designed feature of the readout, not of the world.
• A second clock racing the first. Fatigue against the delivery deadline, where the cure for one costs the other.
• Trade class as a compact price model. Five classes plus a per-node produces/lacks flag generates a whole arbitrage map without a per-commodity price table.
• Spatial portfolio allocation. Presidential Campaign’s weighted winner-take-all states make “where” a harder question than “how much.”
• Coupled indices with decay. Streets of the City’s condition indices all drift the wrong way under inflation, so holding steady already costs money. Deferred maintenance compounds.
• A hard, visible, deterministic clock. Survival’s oxygen is legible every turn, drains only on real action, and never surprises you, which is what makes replaying to beat your own time the actual game.

KN-86 resonance

Concept-level here; the full per-cart mapping across all 17 cartridges is in ../synthesis-basic-games.md.

• The two essays are the most directly actionable material in the entire three-book sweep. They bear on Threshold, on any mission authored as a dependency graph, and on the World Engine’s room/object modeling. Much of the recipe the deck already implements; the value is in what it names that the specs leave implicit.
• Trucker maps onto the two-phase logistics shape the deck already has a cart for, with its speed dial and fuzzy gauge as the tick-loop content.
• Star Merchant and Presidential Campaign extend the economy lineage past the first book’s Hammurabi and Stock Market into spatial arbitrage and weighted portfolio allocation respectively.
• Survival’s deterministic resource race fits the roguelike model and the World Clock with almost no new primitives, and its per-turn status block is a Row-0 pattern.
• Streets of the City is a slower management shape than any current cart, closest to a sector-steward mode.

Assigned to carts

Josh’s calls, made in the kn86-inspo workbench. Full map: cart-inspiration-map.md.

Game	Destination
How To Write An Adventure (essay, 1984)	IF Creation Kit / propaganda, Threshold
Presidential Campaign (1984)	IF Creation Kit / propaganda
Star Merchant (1984)	SynthFence
Streets of the City (1984)	TRADECON
Survival (1984)	NIGHTOWL
Tips for Playing Adventure Games (essay, 1984)	Threshold
Trucker (1984)	Pathfinder

Sources & verification

• Confidence: HIGH on the essays and on the mechanics of Trucker, Star Merchant, Presidential Campaign, Streets of the City, and Survival. All were read from the printed pages, and the essays were additionally OCR’d from 1200 dpi renders so their wording could be quoted accurately.
• Book identity confirmed from the copyright page: Big Computer Games, copyright 1984 by Creative Computing Press, Morris Plains, New Jersey; ISBN 0-916688-40-2; LCCN 84-12730; edited by David H. Ahl. The contents page lists 12 games plus the two adventure essays and a videodisc-adventure article.
• The scan has no text layer. Every page in this capture was read as an image render, and the essays were OCR’d. Quoted phrasing from the essays is reliable; exact BASIC line numbers were read where legible and are cited more sparingly than in the 1978 and 1979 entries.
• Long listings were skimmed, not traced. Presidential Campaign, Star Merchant, Streets of the City, and Trucker each run many pages. Their models above come from the prose descriptions and sample runs corroborated by targeted reads of the listing; per-line arithmetic was not verified.
• Star Merchant’s static pricing is a real property of the listing, not an omission in this reading: a single trade does not visibly move the local price. The prose notes only that a star’s trade classification can drift slightly over time.
• Streets of the City’s index list is taken from the prose description and the sample run’s report table; the exact per-index update coefficients were not extracted.
• Survival’s “no random events” claim is the author’s own, stated in the printed conclusion, and is consistent with the sample run.
• Publication-year note. Some catalogue records place this volume in 1980 alongside the earlier Workman printings. The copyright page of the scanned copy reads 1984, Creative Computing Press, and this entry follows it.