QFlowLearn: QTI 3.0 authoring

Open assessment infrastructure

Assessment infrastructure covers how institutions create, deliver, score, and move digital assessments. This guide explains what each system does and where Question and Test Interoperability (QTI) fits.

The assessment lifecycle

No standard or product covers every step. Institutions need clear system boundaries and records that help staff troubleshoot failures.

  1. Author

    Create semantic items, response declarations, scoring rules, metadata, and accessible content.

  2. Validate

    Check XML, schema, semantics, supported profiles, accessibility, and package safety.

  3. Manage

    Version item banks, permissions, assets, review state, and provenance.

  4. Deliver

    Apply attempt policy, candidate preferences, navigation, identity, and secure item delivery.

  5. Score

    Evaluate trusted responses against authoritative item rules and preserve attempt state.

  6. Report

    Return outcomes to instructors, gradebooks, analytics, and institutional records.

  7. Move

    Export content, assets, metadata, and test records without depending on one platform.

Assessment work moves from authoring through validation, item-bank management, delivery, scoring, reporting, and transfer. Reviews may send an item back to authoring at any stage.

Where QTI stops

QTI describes assessment content and behavior. An institution still needs explicit contracts for launch, identity, roster, accommodations, security, attempts, results, and operations.

Assessment infrastructure boundaries
BoundaryInstitutional responsibility
LMS and launchLearning Tools Interoperability (LTI) or another launch method passes context and roles. QTI carries assessment content and behavior; it does not replace the LMS shell.
Identity and rosterIdentity providers and OneRoster or student information system (SIS) integrations establish people, courses, classes, and enrollments. Candidate records do not belong in portable item content.
Grade returnThe host assessment service decides which trusted outcomes return to the gradebook and how retries, overrides, and scaling work.
AccessibilityAuthors create accessible content. Delivery systems apply preferences and provide keyboard, focus, names, states, messages, contrast, and time behavior.
SecurityHosts own authorization, attempt policy, authoritative scoring, archive safety, asset access, audit records, and incident response.

Guides and research

Choose the guide that matches your current work. Each one cites official sources, public code, known limits, and a checklist.

Compare hosting responsibilities

Hosting location does not determine openness. Evaluate who can inspect the code, export the content, operate the service, apply security updates, restore data, and move to another provider.

  • Hosted: review tenancy, regions, subprocessors, backups, incident response, update policy, data return, and exit support.
  • On-premises: review staffing, release cadence, secrets, observability, capacity, disaster recovery, and the support boundary.
  • Either model: require portable content, reproducible configuration, documented dependencies, and tested recovery.

Institutional evaluation

Test the product with representative content and your institution’s integrated workflows. Record four kinds of evidence separately:

Standard
What the published specification requires.
Implementation
What public or reviewable code provides.
Interoperability
What happened across named independent systems and fixtures.
Certification
What an official program lists for a named product, version, and role.

Where QFlow fits

QFlow is Longsight’s QTI authoring product. Its workflow and the public qti3 packages show how Longsight implemented parts of these guides. Refer to the official specifications for requirements, and test interoperability and certification separately.