Implement RMP
A practical adoption path for institutions, learning platforms, assessment systems, and credential ecosystems.
RMP is an evidentiary layer. It is designed to connect learning environments, assessment systems, adaptive engines, learner records, and credentials—not require their wholesale replacement.
Adoption path
- Define the construct. State precisely what capability or development the implementation intends to represent.
- Inventory the evidence. Separate direct observations, derived measures, inferences, predictions, and contextual signals.
- Choose conformance profiles. Claim only the modules the implementation can actually support.
- Map existing standards. Reuse established competency, assessment, event, rostering, learner-record, and credential standards where they already solve the problem.
- Validate for intended use. Match evidence strength and human oversight to the consequence of the decision.
- Protect rights. Implement consent, access, correction, export, explanation, appeal, retention, and deletion controls.
- Prove portability. Ensure evidence can be inspected outside the system that produced it.
Conformance profiles
RMP-C · Core
Required semantics, provenance, evidence classes, rights, prohibited uses, versioned schemas, and auditability.
RMP-T · Temporal
High-integrity timestamping, synchronization, sequence preservation, sampling information, and latency metadata.
RMP-P · Performance
Direct learning and performance evidence tied to defined constructs and conditions.
RMP-S · Sensor
Optional physiological, gaze, audio, visual, or multimodal signals with explicit consent, quality, and inference limits.
RMP-I · Inference
Model registry, versioning, input provenance, validation, intended use, uncertainty, failure modes, and drift monitoring.
RMP-M · Merit
Merit Evidence Records and assertions with traceable support for capability, growth, retention, transfer, and independence.
RMP-V · Verifiable
Portable, cryptographically verifiable claims or credentials.
RMP-H · High-Stakes
Additional validation, human review, explanation, fairness evaluation, appeals, audit trails, and named accountability.
Interoperability
RMP should map to existing standards where appropriate, including 1EdTech CASE, QTI, Caliper Analytics, OneRoster, Comprehensive Learner Record, Open Badges, and W3C Verifiable Credentials. RMP adds evidence provenance, temporal structure, assistance state, longitudinal trajectory, confidence, and merit assertions.
The decision rule
The weaker the evidence and the higher the consequence, the less authority the automated inference should have.
To discuss a pilot, adoption profile, or integration, use Request a conversation in the navigation.