Arc

The theme registry defines Aurora, Aurora Light, Neon Hacker, Magic Forest and Space. A validated student preference selects the experience; theme resolution has a safe fallback. The learning records do not need to be duplicated for each visual style.

The trade-off. Themes change navigation, progress treatment and atmosphere. They are presentation choices, not separate games or evidence of improved learning outcomes.

The task registry dispatches to specialized handlers. Canvas and flowchart work export visual artifacts; matching stores structured answers; programming has a Browser Coder boundary; shared sessions finalize participation into submissions. Reflection, sorting, questionnaires and external launches have their own paths.

The trade-off. The canonical enum lists 21 kinds, including games and instructor-led or external activities. That is not a claim of 21 equally complete standalone editors. The films exercise a selected set of native activities.

Submission services distinguish draft updates from submitted work. A late autosave does not silently invent a new draft after submission. Published reviews belong to the submitted version, so later student changes do not erase the context the instructor reviewed.

The trade-off. Each editor must respect that lifecycle. Text revision is demonstrated; the captured flowchart’s new-version editor did not automatically copy the previous editable diagram, so that behavior is not promised here.

Quiz services validate the current question, membership, active state and timing before accepting an answer. Instructor controls close, advance and end rounds. Other shared modes have distinct rules: moderated word clouds, anonymous-position heatmaps, turn-based debate and branching votes.

The trade-off. Realtime presentation needs a recovery path. The demonstration uses Arc’s polling fallback with independent browser sessions; it does not establish WebSocket latency or production performance under load.

The smart draft generator scores existing activities using tags and preferences. An author can edit the result and inspect coverage. Publishing creates a reusable pack; assigning a pack to a class and releasing its lessons are separate operations.

The trade-off. Preferences influence ranking rather than guaranteeing exclusive task types or hard tag constraints. The captured draft deliberately contains preparation gaps; generation is not proof that a curriculum is ready to teach.

Analytics intersects actor restrictions with the requested reporting scope, then includes that context in cache identity. School views can connect aggregate activity to permitted classrooms and submission previews.

The trade-off. Participation and learning-friction indicators describe recorded activity. They are not validated educational outcomes, and questionnaire profile labels are not a clinical assessment.

This portfolio demo presents selected read-only states for each catalogued capability. Its charts, filters and examples use a local fictional dataset. It does not connect to learner accounts, run code, submit work or call Arc’s services.

Original screenshots and recordings were captured from an isolated copy of Arc with synthetic accounts. They document real application behavior at their recorded revisions. Source review, the portfolio demo and production operation are distinct evidence.