Initial design notes: critical questions

[alice-i-cecile]

Emergence Design Notes

(Brian, Rose, Alice)

Project summary

Colony of symbiotic insects.
Symbiotic with plants, fungi and so on: permaculture theming.
Speciation and genetic control.

Organic hex-based factory builder.
Indirect control early.
2.5D; height but no overhangs.
Players encouraged to make multi-nuclei bases.
Strong spatial variability: stronger prompts, less blueprint spam.
Strong temporal variability: encourage automatic robustness.
Strong environmental control.
Units have their own needs that must be fulfilled.

Conservation laws: energy, mass, elemental identity.

Workers, not belts and inserters.
Share resources via mycorrhizae (power pole analogue).

Programming game, but with hyperparameters instead of code.

Critical questions

Can indirect-only control still feel tight enough?

Tools:

1. Zoning
2. Species-specific settings
3. Pheromones

Can we set expectations well?
How do we provide high-quality debug information when problems arise?
How do we avoid death-spirals?

Concerns:

• frustrating
• slow to pivot

How does signal-propagation scale?

1. Everything emits a signal (varies temporally and based on quantity)
2. Signals spread to neigboring tiles
3. Signals decay over time
4. Units use signals to determine what they're going to do (combined with preference weight matrix)
5. Units move up the signal gradient until reaching the source

Scales very well: computation is only done once no matter how many units we have.

Needs benchmarking to make sure order of magnitude.
10k tiles to a side, 1k ants at minimum.

What are the performance characteristics for each variable?

Tangential questions

Ants, or not-ants?

How do we evaluate this question?

Real-world species pushes us into a simulationist perspective: harder to deviate.

Notes

• spatial variability
  • e.g. climatic zones, weather?
  • "plan based gameplay" -> go in with a plan, but details are left unknown due to randomness
• temporal variability: gluts/famines of resources, or disruptions to established routines
• hooks
  • a single method on the public API of a system
  • a part of a mechanical subsystem that can be used by other mechanical subsystems
  • enables content generation without exploding complexity budget
• first law of geography: things that are close together are more similar
• tuning levers
  • property of the system that can be adjusted
  • good engineering patterns: orthogonality, grouping, valid over entire range...
  • can be tweaked easily
• constraint
  • things you must / can't do with your design
  • make it easier to design