- Challenges to assumptions — if a number is wrong, show why with a source
- Better data — ONS, IFS, IEA, academic papers preferred over political sources
- Alternative parameter sets — "what if LVT is capped at 80%?" is a valid contribution
- Historical analogues — what happened when Estonia, Singapore, Pennsylvania tried similar things
- Identified weaknesses — things the model does not account for
- Policy based on who deserves what — that is philosophy, not economics
- Keyword lists or heuristics substituting for data
- Contributions that serve the goals of any existing political party rather than the five goals in README.md
- Open an issue first for anything significant — describe the problem before the fix
- Fork, branch, change, open a PR
- Every changed parameter needs a source linked in
docs/sources.md - Every new claim in the manifesto needs a corresponding line in the model
The model (model/) is value-neutral. It runs numbers. It does not argue. The manifesto (manifesto/) draws conclusions from the model. Keep them separate. An economist who disagrees with Swinkism politically should still be able to run the model and agree that the arithmetic is correct.
The critique/ directory is not a graveyard — it is a live record of the strongest objections. If you find an objection that is not in there, add it. If you think an existing critique is wrong, argue it in the issue, not by deleting it.