Papers

LawVM’s public research material explains the compiler model, benchmark limits, and evidence vocabulary. The first paper is in preparation.

Paper in preparation

Finland replay compiler in preparation

Working title: Reconstructing Point-in-Time Statutory State from Amendment Streams: A Replay Compiler for Finnish Law

LawVM replays Finnish amendment streams into point-in-time text-state, compares the result with a declared consolidation witness, and classifies residual mismatches instead of collapsing them into one error bucket. An August 2026 development run reported 0.23% mean normalized text edit distance. That result is project-reported, not an error or correctness percentage.

Status: drafting in progress. The public evidence ledger reports the Finnish candidate work only in aggregate while external disposition is pending.

Public wording: the current assurance pages, evidence records, and dated results define the public claim ceiling. Draft paper language stays within that ceiling.

Related public research

The essays and technical-notes index provides the complete reading path, including the civic and structural essays.

• Parsing amendment language — prior work, deterministic grammar, and recovery limits.
• Reading benchmark numbers — what comparison metrics do and do not establish.
• Truth surfaces in legal information systems — enacted, published, revised, and replayed representations.
• Replay bugs versus editorial artifacts — typed divergence classes and alternative explanations.
• Corrigenda as operations over operations — correction provenance and temporal consequences.

For the current metric and method, see Artifacts. For bounded mechanism claims, see the assurance overview.

Citation

Until a paper is published, cite the software:

Kunnas, E. (2026). LawVM: An open-source replay compiler
for amendment-driven law. https://github.com/eliask/lawvm

A CITATION.cff file is included in the repository.