Record the v1.0 audit items in the roadmap

[JUhalt]

Brings the roadmap up to date with the work merged since 0.8.1. Documentation only.

• Audit items. The v1.0.0 section gains "Audit before the release": the fourteen items from the October 3, 2026 audit, each with its issue and, where merged, its pull request. Eleven are done ([v1.0] Estimate and report the pretest (testing) effect; make the pretest coefficient interpretable #104–[v1.0] Reproduce the Walton Braver and Braver (1988) worked example; correct the roadmap's validation claim #109 and [v1.0] report_solomon() fails on several-treatment summary fits; test every result class #111–[v1.0] Make Satterthwaite the default inference of fit_solomon_ml() #115, merged in Report the pretest effect, name link scales, and draw TOST intervals (#104, #114, #107) #121–Make Satterthwaite the default inference of fit_solomon_ml() (#115) #124 and Credit the published sources: Newman et al. (1990), Williams & Newman (1982), Dukes et al. (1995), and the 1988 worked example (#105, #106, #109) #126). Three are open: the robustness study ([v1.0] Robustness study of the recommended inference before tagging #116), one output vocabulary ([v1.0] Freeze one output vocabulary across fit classes #110), and no experimental functions ([v1.0] No experimental functions at 1.0: implement each one to its literature and meet its validation rule #117).
• Sources Steyn (2009) names. [v1.0] Sources named by Steyn (2009): Kerlinger, Steyn & Mynhardt, Steyn (2001), Scheffé (1953) #96 is now ticked, with its pull requests (Credit Scheffé (1953) and cite Steyn and Mynhardt (2008) (#96) #102, Fix the win-builder NOTE (#84); record the remaining Steyn sources (#96) #103) and a note that Kerlinger's textbook and Steyn's (2001) thesis are recorded as named but not read.
• Bayesian methods. The "Later" list links [Post-1.0] Bayesian analysis of Solomon designs #118 and the Post-1.0 milestone.

🤖 Generated with Claude Code