Repository navigation
Glossary
Philip Pavlik edited this page Oct 9, 2026
·
11 revisions
Last Updated: 2026-10-09 | For: Students, Teachers, Admins, Developers
This page is the canonical source for wiki terminology.
| Term | Canonical Definition |
|---|---|
| Lesson | The learner-facing content item shown on the dashboard. In authoring/upload terms, a lesson is the uploaded TDF package (TDF + stimulus file + optional assets such as images/audio/video). |
| TutorScript | JSON authoring language for lesson packages, including SPARC documents and production rules. |
| SPARC document | Declarative instructional page with stable nodes, learner-action mappings, working-memory facts, and executable rules. |
| Production rule | Conditions and optional tests over working-memory facts, followed by effects such as feedback, model practice, document changes, or an instructional decision. |
| Knowledge component (KC) | Author-defined concept or skill identity used to group learning evidence. |
| TDF | Tutor Definition File. JSON configuration that defines lesson structure and behavior (tutor.setspec, tutor.unit, deliverySettings, session objects, etc.). |
| TDF package | The ZIP upload bundle containing at minimum a TDF and its referenced stimulus file, plus optional assets. |
| Root TDF | The primary TDF entry for a lesson that can include randomized condition variants (sub-TDFs). |
| Sub-TDF | A condition variant under a root TDF, typically tracked for assignment/counts rather than shown as a separate lesson. |
| Term | Canonical Definition |
|---|---|
| Stimulus file | JSON content file referenced by setspec.stimulusfile; stores the instructional items/questions and answer data. |
| Canonical media reference | The expected asset identifier/reference generated by upload/conversion flows; authoring should target canonical references, not non-canonical filename recovery behavior. |
| Cluster | A group of one or more related stimuli (stims) in the stimulus file. Many scheduling rules select at the cluster level (for example clusterlist). |
| Stim / stims | A single stimulus entry (stim) or the array of stimulus entries (stims) inside a cluster. A stim contains fields like display and response. |
| Schedule | The generated per-student sequence of assessment trials produced from an assessmentsession template. |
| Group template | In assessment sessions, the conditiontemplatesbygroup pattern that defines how clusters repeat and are positioned. |
| AutoTutor session | Conversational unit configured with autotutorsession; reads an authored autoTutor script from the first stim in its configured cluster. |
| H5P historical records | Retained records from the removed H5P display runtime; new H5P authoring/upload/playback is unsupported. |
| Condition-root export | A report/export scope that can include multiple related condition branches for a root lesson target. |
| Unit template | A reusable unit definition stored in setspec.unitTemplate, commonly used by adaptive video logic. |
| Term | Canonical Definition |
|---|---|
deliverySettings |
Unified learner-runtime settings for timing, sequencing, feedback, answer controls, prompt presentation, and related lesson behavior. It can appear at tutor.deliverySettings, tutor.unit[].deliverySettings, and tutor.setspec.unitTemplate[].deliverySettings. |
| Legacy delivery setting names | Older content and historical docs may use deliveryparams and uiSettings. New authoring should use deliverySettings. |
d(x, y) |
Data-export helper notation meaning: use x if present; otherwise use explicit alternate value y. |
| Term | Canonical Definition |
|---|---|
| Embedded API keys | Lesson-embedded speechAPIKey / textToSpeechAPIKey from TDF setspec; used directly when route/config enables embedded-key mode. |
| Account menu | The top-right authenticated user menu that shows avatar, display name, role, and user actions such as Profile, Audio settings, theme selection, and Log out. |
| Profile | User settings page reached from the Account menu; stores private account name, public display name/avatar, and personal OpenRouter settings. |
| Profile flow | Standard logged-in learner flow where speech recognition normally uses the learner personal speech key (Meteor.user().speechAPIKey) unless embedded-key mode is enabled. |
| OpenRouter profile settings | Personal OpenRouter API key and default model managed from Profile for AutoTutor and AI-assisted content creation workflows. The key is stored encrypted on the MoFaCTS server and returned to the browser for client-side OpenRouter requests. |
| Experiment logins | Login/entry routes that enable embedded-key mode (useEmbeddedAPIKeys = true) so lesson-embedded keys are used directly. |
Getting Started
- Teacher: First Hour
- Quick Start: Content Creation
- Class Setup and Assignment Workflow
- Anki Import Guide
Authoring Reference
- Introduction to TutorScript
- Content Creation Reference Tables
- Stimulus Files (Content)
- Trial Types Reference
- Audio and Speech Settings
Advanced Authoring
Reports and Help
- Researcher Guide
- TDF field tables
- Gaze tracking prototype
- Data Output and Reports
- Data Privacy, IRB, and FERPA Notes
- Experiment Routes
- SPARC Instructional Cycles
- Public Demonstrations
- Learning Algorithms Reference
- Video Units and Adaptive Logic
- AutoTutor Sessions
- Deployment Guide
- Security Audit Operations
- Remote Install
- Runtime Settings Reference
- Admin Features Reference
- Custom Help Page Setup
- Troubleshooting
- Local Install
- Developer Reference Guide
- Learning Algorithms Reference
- Introduction to TutorScript
- AutoTutor Sessions
- Runtime Settings Reference
- Deployment Guide
- Glossary
- FAQ
- Screenshots
- License Guidance
- Implementation Cost Analysis
- GitHub Repository
- GitHub Issues