Design-Research Frameworkv0.1 · working draft

A working framework for product and UX leads

How to think about the whole of UX work, with agents in it.

Five domains in tension, three loops that feed each other, and one store every cycle adds to. Kept since 2011, published here as a working draft so you can read it, copy it and build your own harness from it.

The three loops of the Design-Research FrameworkThree linked loops: Strategy (Direct and Define), Envisioning (Discover and Detail) and Making (Develop and Deploy). All three feed the well-informed space beneath them.STRATEGYDirectDefineENVISIONINGDiscoverDetailMAKINGDevelopDeployTHE WELL-INFORMED SPACE

Pillar one

The framework

  • The problem space

    Five domains that pull against each other. The tension between them is where product problems come from.

  • The three loops

    Discovery and design, develop and deploy, direct and define. Three loops that feed each other, with the 6Ds as their phases.

  • The well-informed space

    The store every cycle adds to. Observations, articulations, models and reports, structured so people and agents can find and build on them.

  • Cost of contact

    No-touch, low-touch, high-touch. The rule for which research an agent can run and which needs a person in the room.

  • The Triton method

    Dev-led, triple-track. Development ships first, UI design iterates on what's built, and research evaluates what shipped.

Pillar two

Artifacts

Each activity produces one defined output, and each output feeds the next. In a harness, every artifact is the output of a skill, with a shape the next skill can read. The chain below is the 2010 spine.

  1. Raw data
  2. Matrices
  3. Personae and models
  4. Requirements
  5. Approved design proposal
  6. Form and behavior spec
  7. Reports

For agents

This site is a well-informed space

Every page is a Markdown file with the same front matter. People read the pages. Agents read the same records through llms.txt and feed.json, and every page links to its source.

---
title: Problem articulation
type: artifact
loop_stage: discovery/interpreting
domains: [people, business, technology]
status: draft
version: 0.1
sources: [DRF 2010 spine]
confidence: stated
---

Record

Origins

One dated timeline from 2010 to now, and the pages first published here in February 2024, kept unchanged. See the record.