Taxes and science

Each civilization has a treasury (money), a science rate (sci) and a tax rate (civ_toll). Rates are in tenths. Luxury spending is the remainder, 10 - sci - civ_toll. civ_research_tally stores research progress; spaceship_arrive_dates stores spaceship arrival years.

CITYMGT.C calculates city trade, corruption, route income, taxes, luxuries and research. ECONOMY.C processes the cities and adjusts AI rates, chooses research and records discoveries.

Corruption

The city pass finds distance to an owned Palace, starting with a maximum of 32. A city with a Palace starts at distance zero. Communism replaces the distance with 10. For government ID g, the loss is:

loss = trade * palace_distance * 3 / (g * 20 + 80)

A Courthouse halves it; Democracy sets it to zero. The calculation is run before and after route income is added. This is a trade deduction; the source does not apply this formula to shields.

Splitting trade

After corruption and route income, luxuries are rounded with +5 before dividing by ten. Taxes are rounded the same way and limited to the trade left over. Science receives the remainder. Each specialist then adds two to its chosen category.

A Marketplace adds half the current taxes and luxuries. A Bank then adds half of those updated amounts, so the bonuses compound. A Library and a University each add half the original science amount. Wonders can modify science further.

During turn processing, a city in order adds its taxes unless its government is Anarchy. Building maintenance is charged later in city_process(). Military support is deducted from city shields, not paid as a general one-gold-per-unit treasury charge.

Research

give_advance() records discoveries in advances_known, updates discovery counts and sources, and handles effects such as obsolete Barracks and Great Library transfers. has_tech() tests the packed bits. Technology prerequisites and names come from advances[72] in GAME.C.

Research thresholds depend on advances already known, difficulty, the calendar and other modifiers in city_process(). The human research selection is selected_index; the AI chooses with weights from advance_value(), including the leader’s personality.