Compare
Side-by-side comparisons
Improved samples stacked side-by-side — findings, hardening cost, and when each choice fits.
A comparison here is not a feature checklist copied from documentation. Each option is run in Docker, the same production-hardening work is applied to every one of them, and what gets tabulated is what came out: the findings that actually appeared, the number of commits each fix took, and the test counts before and after.
That means the columns are commensurable. When two options show the same hardening cost, it is because the same work was done to both and measured the same way — not because they felt similar.
How to read one
- Findings are what the verification turned up on that option specifically. An empty cell means the problem did not occur there, not that it went unchecked.
- Hardening cost is measured in commits and test deltas, both reproducible from the published fork.
- When it fits is the only editorial column, and it says which decision the measurements support rather than which we prefer.
Every comparison names the versions it was verified against and the date. Findings are true as of that date; when we re-check an older one, the result is added to it whether or not anything changed.