Bound the claim.
State the consequence, world, evidence, and action cut before treating a result as portable.
About RISU Institute
Semantic assurance for machine action.
RISU Institute is an independent research initiative developing carrier-neutral methods and executable infrastructure for consequential machine action.
Research purpose
Make the consequence-relevant distinction explicit before a machine acts.
Many software boundaries preserve names, fields, and apparent compatibility while silently changing what another system may infer, rely on, or cause. RISU studies those load-bearing distinctions directly: what a machine-readable claim means, what evidence supports it, when another system may rely on it, and whether the distinctions that determine an action's consequence are still operative at the effect.
The objective is not to build a vocabulary for one protocol or vendor. The objective is to identify small, reusable semantic objects and verification conditions that remain meaningful across carriers.
Research method
Conceptual distinctions become technical obligations only after they survive pressure.
Several RISU projects begin from distinctions with long histories in legal and institutional reasoning—authority, revocation, reliance, recognition, absence, finality, notice, supersession, evidentiary status, and operative effect. They are used as sources of technical design vocabulary, not as legal conclusions.
Existing standards, formal literature, and deployed systems define the starting point. A new abstraction is introduced only where a bounded operational gap remains. The abstraction is then attacked with adversarial cases, constructive controls, source conformance, differential checks, mutation tests, or real-system commissioning according to the claim being made.
Preserve the evidence.
Keep direct observation, source-derived fact, analytical premise, and formal consequence distinguishable in the public record.
Attack the mechanism.
Use counterexamples and hostile controls to test whether the assurance result depends on the intended mechanism rather than a surface proxy.
Keep the layer neutral.
Use particular APIs, protocols, or products as test carriers without allowing them to define the semantic substrate.
Publication discipline
The public record should show the boundary, not hide it.
Published claims are scoped to the evidence that supports them. Experimental status, assumptions, trust boundaries, failed mechanisms, and unresolved questions remain visible. Model-relative reasoning, source preservation, and operative system evidence are not treated as interchangeable achievements.
RISU Technical Notes provide stable scholarly records. Supporting code, datasets, certificates, and archived releases are linked where they materially support the publication. Peer review, independent reproduction, ecosystem adoption, and production deployment are never implied by publication alone.
Organization
Independent by design; public by default where the claim permits it.
RISU Institute is a research initiative of RISU LLC. The institute's public role is to develop and publish research methods, technical notes, executable verification artifacts, and research software in a way that lets readers inspect the relationship between a claim and the evidence supporting it.
Public source and experimental artifacts are maintained through the RISU Research GitHub organization.