Build plan
R4PC — First build roadmap
A pragmatic path from current prototype to a public, contributable platform — scoped to the next ~6 months.
This roadmap covers the next six months of build work, scoped to a small founding team plus part-time contributors. It is intentionally conservative: ship narrow, prove the method on a real case, then widen.
Where we are today (M0 — done)
- Standalone interactive HTML prototype (Punjab flagship case).
- Master strategy document (v1).
- Platform blueprint (v2) defining method, modules, governance.
- Brand identity: R4PC wordmark + 4Ps Venn (navy / gold / paper).
- Public landing page (this site) — independent R4PC framing.
- Clean Git repo, ready for GitHub.
Milestone 1 — Public preview & narrative lock (Weeks 0–2)
Goal: anyone you send a link to understands what R4PC is, what method it uses, and what stage it's at, in under 90 seconds.
| # | Task | Done when |
|---|---|---|
| 1.1 | Push this repo to GitHub under the R4PC org / personal account. | Repo public, README rendering. |
| 1.2 | Wire a preview domain (Cloudflare Pages or GitHub Pages → r4pc.org or staging subdomain). | HTTPS site live at chosen URL. |
| 1.3 | Tighten OG / canonical URLs in index.html to the chosen domain. | og:url + <link rel=canonical> correct. |
| 1.4 | Add LICENSE (code: MIT; content: CC BY 4.0 recommended). | Both files in repo root. |
| 1.5 | Run a 4-person comprehension test (1 policy person, 1 researcher, 1 funder-adjacent, 1 lay reader). | Notes archived; copy edits applied. |
Exit criteria: site lives at a real URL; 4 testers each correctly describe (a) what R4PC is, (b) what the 4Ps mean, (c) why Punjab is the flagship.
Milestone 2 — Methods Hub v0 (Weeks 2–6)
Goal: publish the 4Ps method as an open, auditable specification. Not a product yet — a document set.
| # | Task | Done when |
|---|---|---|
| 2.1 | Write methods/4Ps-overview.md — definitions, scoring rubric, intended uses & limits. | Reviewed by 2 external readers. |
| 2.2 | Write methods/case-template.md — the canonical structure every R4PC case follows. | Used to retrofit Punjab case below. |
| 2.3 | Move methods/ into the website as a static section. | Linked from the main nav. |
| 2.4 | Open issue templates on GitHub: "Submit a problem", "Submit a case", "Suggest a methods change". | Templates merged. |
Exit criteria: an external researcher can read the Methods Hub and replicate a 4Ps evaluation on their own case without contacting R4PC.
Milestone 3 — Punjab flagship as a public, web-native case (Weeks 4–10)
Goal: the standalone prototype becomes the canonical worked example, restructured to follow the case template from M2.
| # | Task | Done when |
|---|---|---|
| 3.1 | Rewrite the Punjab prototype as a structured case page (/cases/punjab-crop-residue). | Page live; passes a11y scan. |
| 3.2 | Replace the prototype's data illustrations with static GIS exhibits (PNG/SVG snapshots). | At least 3 exhibits, sourced and credited. |
| 3.3 | Document the 4Ps scoring for Punjab — explicit numbers, evidence, dissent notes. | Section visible on case page. |
| 3.4 | Publish a 1-page "policy brief" PDF derived from the case. | PDF lives at /cases/punjab-crop-residue/brief.pdf. |
Exit criteria: the Punjab page can be sent to a state-government desk, a researcher, or a journalist and stand on its own.
Milestone 4 — Submission Workspace v0 (Weeks 8–14)
Goal: anyone can submit a documented problem or candidate solution into a triage queue. No login wall yet — minimum viable intake.
| # | Task | Done when |
|---|---|---|
| 4.1 | Pick the lightest stack that works: a Google Form / Tally / Formspree → Notion or Airtable backend. | One submission flow from public page → reviewable database. |
| 4.2 | Write a public review SLA (e.g., "we acknowledge within 7 days; we triage within 30 days"). | Published on /submit. |
| 4.3 | Define the triage rubric. | Internal docs/triage-rubric.md. |
| 4.4 | Process the first 5 submissions end-to-end — even if the answer is "not a fit, here's why." | 5 review notes archived. |
Exit criteria: R4PC has received and answered 5 real submissions from outside the founding team.
Milestone 5 — Policy Library v0 (Weeks 12–18)
Goal: a small, curated, searchable library of existing policies and pilots — tagged by sector, geography, and 4Ps outcome.
| # | Task | Done when |
|---|---|---|
| 5.1 | Define the schema (sector, geography, instrument type, 4Ps tags, source URL, year, evidence quality). | Schema in docs/library-schema.md. |
| 5.2 | Seed with 30 entries (start with air quality + agriculture). | 30 entries live, each with a source link. |
| 5.3 | Implement search/filter on the static site (client-side JSON). | Page works without JS for read-only fallback. |
| 5.4 | Open contributions via a structured PR template. | First external PR merged. |
Exit criteria: a practitioner can land on the Library, filter by sector + geography, and pull 3 useful prior policies in under 5 minutes.
Milestone 6 — GIS Lab v0 (Weeks 16–24)
Goal: at least one second case (beyond Punjab) is published with a real GIS validation step, using public datasets.
| # | Task | Done when |
|---|---|---|
| 6.1 | Pick the second case (candidates: California groundwater / land repurposing; urban heat). | Case selected, scoping doc published. |
| 6.2 | Run the case through the case template, including GIS validation. | Static exhibits and methodology written up. |
| 6.3 | Document the GIS toolchain (QGIS / kepler.gl / public layers). | docs/gis-toolchain.md. |
| 6.4 | Invite one external GIS contributor to audit and replicate. | External audit note published. |
Exit criteria: R4PC is no longer a single-case demo — the method has been run end-to-end on two distinct cases.
Deferred (intentionally not in this roadmap)
- AI Navigator — only after Methods Hub + Library v0 exist.
- User accounts / auth — not until submission volume justifies it.
- Custom backend — stay on managed services (Notion / Airtable / static JSON) until a backend pays for itself.
- Mobile app — never, unless a specific field-data use case demands it.
Operating principles
- Ship narrow, finished work. A small, finished module beats a large, half-built one.
- The method is the product. R4PC's defensibility is the 4Ps + GIS validation discipline, not any particular piece of software.
- Independence is non-negotiable. R4PC is its own platform.
- Public by default. Cases, methods, and decisions are published unless there is a specific reason not to.