← All documents

Research

Background that informs the design. Useful as input, not as authority.

What's here

lore.md Read first. Fiction as design source, and why
what-fiction-skips.md The problems no story has to solve — where original design lives
archetypes.md Twelve field-operator archetypes for stress-testing designs
ecosystem-roster.md Watch, analysts, agents, infrastructure, outside, adversaries
constraints.md What those stresses demand of any feature
prior-art.md What already exists, and what it teaches

How to use the archetypes

They're a stress-testing device, constructed from published material about real-life superhero and volunteer street outreach communities — not interviews, not evidence about real people.

The correct use is: design a feature, then check it against the twelve to find what breaks. Breakage tells you what the feature needs — a visibility setting, an offline path, a smaller footprint, clearer semantics.

The incorrect use is to treat a failure as a verdict. The roster is deliberately weighted toward operators who refuse, break or abandon software, because that's what surfaces problems early. A roster weighted that way will reject almost anything if you let it vote. It surfaces requirements; it does not decide scope.

When an archetype refuses a feature, the first question is always "what setting would let them opt out while everyone else keeps it?" Removal is the last resort, not the first.

What real evidence looks like

Everything here is hypothesis until operators use the thing. Real signal comes from:

  • Someone corrects a directory entry without being asked
  • Someone adds knowledge nobody prompted them for
  • Someone opens the app on a night with no op scheduled
  • Someone travelling uses it to work with people they've never met
  • Someone says they'd notice if it disappeared

Those beat any amount of analysis in this folder.

docs/research/README.md