These four documents are dated snapshots from 2026-05-27. They are the
analysis that produced the restart decision and informed the v0 scope. They
are not living docs — the living docs are
../PRD.md and ../ARCHITECTURE.md. Treat
this folder as the project's reasoning archive.
| Doc | What it answers |
|---|---|
assessment.md |
Why restart — vs forking or upstream PRs to doctly/switchboard. The strategy framing (Alt 1 vs Alt 2) and the "fixing quality is more hazardous than building fresh" heuristic that anchors everything else. |
feature-sizing.md |
T-shirt sizes per feature, MoSCoW, and the Continue-vs-Restart estimate (cold rebuild ~86 PD vs lifting MIT modules ~69 PD vs the author's actual ~55 PD). Appendix A is the git-log extrapolation. |
electron-architecture.md |
The as-built architecture of the Electron app (the reference implementation you'll often want to consult during the restart — module map, data flows, the IDE/MCP protocol, the known structural risks). |
electron-nfrs.md |
The Electron app's NFR scorecard — packaging, signing, auto-update, CI, security, performance. Source of truth for "what gap is Q<n> fixing?" cross-references in PRD.md §4. |
assessment.mdpredates the rename to TermHerd and is written from the perspective of evaluatingdoctly/switchboard(which the author did not own). Its conclusions — restart, MIT-lift the domain modules, defer the IDE — are preserved verbatim because they are the reason this repo exists.- The original PRD (
../PRD.md) and architecture (../ARCHITECTURE.md) were also drafted on 2026-05-27 and have since been revised (rev. 2) when the MoSCoW reshuffled the IDE/diff into the Unsure bucket. The numbers infeature-sizing.mdreflect the pre-rev.2 estimate and are still useful as the budget anchor; the post-rev.2 milestone plan lives in../PRD.md§9.