Rich Assessments & Integrations

Overview

Assessment & Grading covers how authors define learning assessments — from simple multiple-choice problems (CAPA) to open-response essays (ORA), drag-and-drop exercises, proctored exams, and staff-graded assignments. It also covers the grading policy framework that determines how component grades roll up to course grades.

This is one of the platform's most complex and historically deep feature areas, with the CAPA (Computer-Assisted Problem Answering) system dating to the original MIT platform (6.002x).

Current State (2026)

• CAPA: The primary problem type framework — supports multiple-choice, checkbox, text input, numerical, custom Python-graded, and many more; authored in XML
• ORA (Open Response Assessment): Multi-step assessment with peer review, staff grading, and rubrics; major standalone system
• Proctoring: Proctored exam configuration via edx-proctoring; supports multiple proctoring vendors via a plugin interface
• Grading policy: Course-level grading rules (assignment types, weights, cutoffs) defined in Studio settings; computed by LMS grading engine
• Staff grading: staff-graded-xblock for assignments where instructors manually enter grades
• Problem authoring: Problems authored in OLX XML within Studio; advanced editors exist for some types

Architecture

• CAPA: XBlock-based problem rendering and grading in edx-platform; Python-graded problems run in codejail sandbox; submissions stored in StudentModule
• ORA: edx-ora2 is a Django app + XBlock; uses edx-submissions for submission storage; separate peer review workflow
• Grading engine: Subsection-level grade computation in edx-platform; grades stored per-learner, per-subsection; recomputed on policy changes
• Proctoring: edx-proctoring provides a provider abstraction; third-party vendors (Proctorio, Honorlock) integrate via the proctoring API

Relevant Repositories

Repository	Role in This Feature	Activity Level	Notes
openedx/openedx-platform	CAPA problems, grading engine, course policy	High	Core assessment
openedx/edx-ora2	Open Response Assessment	Medium	Peer/staff grading
openedx/edx-submissions	Submission storage (backs ORA, CAPA)	Medium	Shared layer
openedx/edx-proctoring	Proctored exam backend	Medium	Vendor abstraction
openedx/xblock-drag-and-drop-v2	Drag-and-drop problem XBlock	Low	Rich problem type
openedx/staff-graded-xblock	Staff-graded assignment XBlock	Low	Manual grading
openedx/frontend-app-ora	ORA learner-facing MFE	Medium	ORA learner UI
openedx/frontend-app-ora-grading	ORA staff grading MFE	Medium	ORA grading UI

Recent Changes

• ORA MFEs (frontend-app-ora, frontend-app-ora-grading) maturing
• CAPA problem type improvements ongoing

History

Origin

• Year introduced: 2012 (CAPA predates Open edX; originates from MIT 6.002x in 2012)
• Initial implementation: CAPA (Computer Aided Problem Answering) system from MIT; ported from the original MIT course infrastructure
• Context: CAPA was MIT's internal problem system; when edX was built, it incorporated CAPA as the foundation for graded problems

Key Milestones

Year	Milestone	Teams / People Involved
2012	CAPA system incorporated from MIT	Unknown
~2013–2014	ORA (Open Response Assessment) launched	Unknown
~2015	Proctored exams introduced	Unknown
~2019–2020	ORA 2.0 with peer review improvements	Unknown
~2022–2024	ORA MFEs development	Unknown

People Who Shaped This Area

• Engineering: Unknown — open question for interview
• Product: Unknown — open question for interview
• Design: Unknown — open question for interview

Open Questions

• [ ] What was the original CAPA system from MIT and how much was preserved vs. rewritten?
• [ ] When was ORA first introduced and what was the academic motivation for peer review?
• [ ] How does the Python-graded problem type work with codejail?
• [ ] What proctoring vendors have been integrated historically?
• [ ] What are the limitations of the CAPA XML authoring format that authors find most painful?
• [ ] How is the grading engine architected — what happens when a grading policy changes?