Skip to content

Roadmap

Jump to bottom
Wouter Meetsma edited this page Jun 24, 2026 · 48 revisions

Development Roadmap

🎯 Overview

RubricMaker is an offline-first rubric creation and grading tool for educators. This roadmap lists completed phases first (newest at the top of that section is the most recently shipped), then planned work.

✅ Completed Phases

Phase 0: MVP & Core Features (Launched)

  • ✅ Rubric builder with multiple scoring modes (Total Points, Weighted, Single-Point)
  • ✅ Interactive grading interface with score modifiers
  • ✅ CEFR framework integration (Can-Do statements)
  • ✅ Student self-assessment & peer review
  • ✅ Essay assignments with rich text editor
  • ✅ Analytics dashboard with class statistics
  • ✅ Multi-format export (PDF, DOCX, CSV)
  • ✅ Student portal with shareable links
  • ✅ Offline-first architecture with localStorage
  • ✅ Basic theme customization

Phase 1: Cloud Sync & Collaboration (PR #128)

  • ✅ Sync conflict resolution — src/utils/syncMerge.ts, remote-as-baseline merge with LWW on rubrics, wired into all three hydration sites (startup, reconnect, OTP sign-in)
  • ✅ Cambridge English framework — CEFR ↔ exam mapping (src/data/cambridgeExams.ts), opt-in "B2 · FCE" badges, optional Cambridge Dictionary API vocabulary lookups
  • ✅ Theme bundles — 5 one-click presets (src/data/themes.ts), tonal accent scale via color-mix(), 5 decorative export fonts for PDF/DOCX

Phase 2: Infrastructure Stabilization (PR #131 + wiki updates)

  • ✅ 2.1 Multi-device e2e — second-browser-context Supabase fixture, propagation/LWW-race/network-partition specs (e2e/specs/18-multi-device-sync.spec.ts)
  • ✅ 2.2 LWW extended to every editable collection (rubrics, grades, peer reviews, classes, students, grade scales, comment snippets/bank, self-assessments, speaking sessions, analysis results); add/delete-only collections (attachments, exportTemplates, favoriteStandards) intentionally left on pending-queue protection only
  • ✅ 2.3 Wiki documentation pass — Cambridge + theming guides added to Features.md, conflict-resolution strategy documented in Supabase-Sync.md

Phase 3: EFL Classroom Enhancements (PR #134–#136)

  • ✅ 3.1 Speaking session recordings — IndexedDB blob store, local-only audio, DB-gated video, optional cloud sync with cascade delete
  • ✅ 3.2 Vocabulary profiling dashboard (/vocabulary) — per-student/class CEFR distribution charts, open CEFR word lists, CSV export
  • ✅ 3.3 Peer review analytics (/peer-analytics/:rubricId) — consistency scoring, leniency bias, inter-rater spread, comment heatmap
  • ✅ 3.4 Full testing environment (/tests, /test/:code) — MC/short-answer/open questions, auto-scoring, class-average adjustment, live proctoring monitor

Phase 3.5: Hot Text Questions & Standalone Essays Workspace (PR #140)

  • ✅ Hot-text question type — src/utils/clozeParse.ts parses {{cloze}} and [[hot-text]] syntax; scoring in testCalc.ts
  • ✅ Standalone /essays workspace (parallel to /tests) — prompt editor, rubric connector, per-student share links, submission-code import, reusable assignment templates (PR #143)

Phase 4: Admin & School Operations (PR #144–#146)

  • ✅ 4.1 Expanded statistics — multi-class Compare view, trend overlay, rule-based insights, track/year filters
  • ✅ 4.2 Activity dashboard (/activity-dashboard) — rubric/test/essay × class grid, assign-from-view, submission count badges
  • ✅ 4.3 Admin hardening — teacher/admin/observer roles + RLS, pg_cron data retention, student soft-delete archive (restore/anonymize), audit log with CSV export
  • ✅ 4.4 SIS roster sync (partial) — CSV upsert + optional roster-removal sync; native Magister API integration still reserved for community contribution
  • ✅ 4.5 Self-hosting ops — docs/SELF_HOSTING_OPS.md, HestiaCP/Virtualmin guides

Phase 5: UI/UX Polish (PR #147–#153)

  • ✅ 5.1 Mobile grading — touch stepper alongside sliders, single-column ComparativeGrading reflow <768px, iPad CSS, mobile e2e
  • ✅ 5.2 WCAG 2.1 AAA — src/utils/contrastCheck.ts + enforcement test, accessible modals with focus trap/restoration, keyboard drag-and-drop, i18n'd skip link
  • ✅ 5.3 Help & documentation — page tours (Joyride) on 12 pages total, contextual tooltips (src/components/ui/HelpPopover.tsx) for score modifiers/LWW sync/proctoring flags
  • ✅ a11y test coverage — RubricBuilder/GradeStudent/ComparativeGrading added to the jest-axe page suite (PR #152, closing the Phase 5.2 follow-up)
  • ✅ Route-shaped skeleton loading for lazy routes and async actions (PR #153)

Phase 6: Strategic Features (PR #157)

Scoped down from the original proposal before implementation — see "Deferred from Phase 6" below for what was cut and why.

  • ✅ 6.1 Student Learning Paths — rule-based (no AI/LLM) rubric recommendations from CEFR-skill cohort gaps, consecutive-low-score intervention flagging, /students/:id/learning-path (src/utils/learningPathAggregator.ts)
  • ✅ 6.2 Rubric Marketplace (school-scoped) — publish/browse/clone/upvote within a school via the existing schools/school_members tables, /marketplace (supabase/migrations/040_rubric_marketplace.sql, src/pages/MarketplacePage.tsx)
  • ✅ 6.3 Integration Ecosystem (CSV import only) — Clever CSV and OneRoster CSV roster-import auto-detection added to the existing CSV import modal; LTI 1.3 LMS integration deferred (see below)
  • ✅ 6.4 Test Summary Export — per-question and per-skill/standard strong/weak breakdown, PDF/DOCX export (single + batch) from the test results page, feature parity with the rubric summary export (src/utils/testSummaryAggregator.ts)
  • ✅ 6.5 Report Card Improvement — consolidated PDF/DOCX report card per student combining rubric grades, standards coverage, learning goals, CEFR overview, and test summary, with section toggles and single/batch export from the Export page (src/utils/reportCardAggregator.ts)

Verification baseline (carried through every phase above): typecheck/lint clean, full unit suite green (1755+ tests), i18n parity across en/nl/fr/de/es, e2e suite green on chromium/firefox/webkit/mobile-chrome (pre-existing/environmental flakes tracked separately, not regressions).

Phase 7: Longitudinal & Historical Insight (PR #159)

  • ✅ 7.1 Multi-year student history — Student.pastClassMemberships recorded on every class change (manual move via Edit Student, and CSV sync-mode re-import transfers), with a confirmation summary before sync-mode CSV mutates anything; trail rendered on the student profile
  • ✅ 7.2 Rubric version diffing — "Compare" button per saved version opens a diff against the current rubric (criteria added/removed/changed, title/weight, level points/labels) (src/utils/rubricVersionDiff.ts, RubricVersionDiffModal.tsx); version history/snapshot/restore already existed
  • ✅ 7.3 Standards coverage gap analysis — class-level aggregator cross-references every linked standard against what's been graded, split into Assessed/Not yet assessed, surfaced on the Activity Dashboard (src/utils/standardsCoverageAggregator.ts, ClassCoverageGapPanel.tsx)

Phase 8: Collaboration & Department Workflow (PR #160)

  • ✅ 8.1 Co-grading & moderation — second-marker grade reuses the peer-review data model (StudentRubric with isPeerReview: true); src/utils/coGradingModerationQueue.ts flags disputes above a configurable point threshold; /moderation lists disputes with a per-criterion delta and keep/accept resolution
  • ✅ 8.2 Department rubric/comment-bank libraries — sharedWithSchool flag on rubrics and comment-bank items, read-only to every teacher in the same school (041_school_sharing.sql, mirrors the Phase 6 marketplace's school_members join pattern)
  • ✅ 8.3 Grading task assignment — batch-assign a class's ungraded submissions for a rubric to a colleague from the Activity Dashboard (GradingTask type, 042_grading_tasks.sql); completion is derived, not a stored flag
  • ✅ 8.4 Manual reordering — drag-and-drop reordering (src/utils/displayOrder.ts, @hello-pangea/dnd) on RubricList, TestListPage, EssayListPage, the Activity Dashboard, and the class list on the Students page, per teacher. List containers use flexbox-wrap rather than CSS grid, since @hello-pangea/dnd computes drag displacement from sibling bounding rects and doesn't support CSS grid track layout — grid caused incorrect cross-row drag animation and made it impossible to drop a card between two others in a different row.
  • ✅ 8.5 Cohort-based filtering — year/track cohort filter on the Rubrics, Tests, and Essays lists (src/utils/cohortAggregator.ts, src/components/CohortFilter.tsx); a student counts as in-cohort via their current class or any past class from pastClassMemberships, so an item stays visible to a cohort across a class transfer. Manual reordering is disabled while a cohort filter is active. Not added to the Activity Dashboard, since its existing class-level year/track filter already serves the same need for that per-class matrix view.

Phase 9: Accessibility & Reach Expansion (9.1–9.4, complete)

  • ✅ 9.1 RTL readiness — CSS logical-properties audit (src/index.css): physical left/right/margin-left/padding-left/border-right/text-align: left rules converted to inset-inline-*/margin-inline-*/padding-inline-*/border-inline-*/text-align: start, a no-op in the 5 existing LTR locales. New src/utils/rtlLanguages.ts + an AppContext effect sets dir="rtl" on <html> when an RTL language is active — inert today (no RTL locale shipped yet) but ready for one. The mobile sidebar drawer's transform: translateX() slide animation has no logical-property equivalent and was deliberately left sliding from the left edge only — tracked below as a known gap, not fixed in this pass.
  • ✅ 9.2 Installable offline PWA — vite-plugin-pwa (Workbox) generates the manifest + service worker at build time; registerType: 'prompt' (not autoUpdate) so a new build never silently swaps JS under a mid-session sync — the user gets a native confirm prompt (src/pwa.ts) before reloading. Supabase requests (/rest/, /auth/, /realtime/, /storage/, /functions/ + version segment, matched by path so it covers both hosted and self-hosted Supabase) are explicitly NetworkOnly in the Workbox runtime-caching config — the service worker must never make a failed sync request look like it succeeded. Works under the existing relative base: './' sub-path hosting config. Manifest icons: rubric-icon.svg (sizes: any) plus 192×192/512×512 PNG fallbacks (pwa-192.png/pwa-512.png, rasterized from the same SVG via a one-off Playwright screenshot script — no new dependency) for broader Chrome/Android installability. A dedicated maskable purpose variant is still a follow-up, see Known Issues.
  • ✅ 9.3 Dyslexia-friendly reading mode — new Settings toggle (dyslexiaFriendlyMode) sets --line-height: 1.8 / --letter-spacing: 0.04em app-wide via the same CSS-custom-property pattern as the existing theme bundles; no new component, just two new :root vars in src/index.css with their var(--x, default) fallback as the single source of truth for the off state.
  • ✅ 9.4 Expanded essay exports — new "Essay Export" section on the Export page (src/pages/ExportPage.tsx) and src/utils/essayExport.ts. Exports a TipTap essay submission (EssaySubmission.contentHtml) to Markdown, DOCX, or PDF, either as a single essay, separate files for several selected students, or combined into one document (Markdown --- join, DOCX PageBreak, PDF page-break-after). No new data model was needed: the essay/rubric/analysis combined export (exportEssayWithRubric) joins EssaySubmission ↔ StudentRubric ↔ DocumentAnalysisResult at export time by (rubricId, studentId), rather than adding a stored attachment FK — avoiding the schema change the original proposal assumed was required. The HTML→Markdown/DOCX conversion is a small custom DOM walker scoped to exactly the tags TipTap's EssayEditor emits (no new dependency); tables/task-lists degrade to plain text in Markdown/DOCX (PDF keeps full fidelity since it prints the HTML directly via the existing printHtml/browser-print pattern). Batch "separate files" mode loops single-file downloads rather than zipping — no JSZip added.

🔭 Future Roadmap: Phase 8 and beyond

Each phase below groups proposals under one theme so they can be scoped and shipped independently. None of these involve AI/LLM content generation — that remains explicitly out of scope for this project.

Deferred from Phase 6

Two items from the original Phase 6 proposal were cut from scope before implementation, since neither could be built and verified in the implementation environment:

  • LTI 1.3 LMS integration (Canvas, Blackboard, Moodle) — needs a live LMS sandbox to verify OAuth/launch flows; CSV-based roster import (Clever, OneRoster) shipped instead as a same-outcome, locally-testable substitute for the common "get students into the app" need.
  • Public/cross-tenant rubric marketplace — shipped as school-scoped only (reuses schools/school_members); a public marketplace would need a moderation queue, abuse controls, and report/flag tooling that don't exist yet.

Both remain candidates for a future phase if there's demand.

Phase 9 (Accessibility & Reach Expansion) is now complete — see the "✅ Completed Phases" section above.

Phase 10: Data Portability & School-System Interop

Closes the loop on "data leaves the device" use cases beyond JSON backup.

10.1 Gradebook export templates

  • Per-SIS CSV export presets (beyond the generic CSV in Phase 0) for common Dutch/EU school administration systems, so a Statistics export drops straight into the receiving gradebook's expected columns

10.2 Calendar/due-date sync

  • Export essay/test assignment due dates as a downloadable .ics file — no calendar API integration required, keeps the offline-first model intact

10.3 Read-only public API for institutional reporting

  • Optional, opt-in HTTP endpoint (Supabase Edge Function) exposing aggregated, anonymized class statistics for school-level reporting dashboards — gated behind the existing admin role, off by default

📌 Known Issues & Technical Debt

Issue Impact Owner Priority
Native Magister SIS API integration (bulk import + live sync) not built — CSV path covers production use today School IT can't fully automate roster sync Reserved for community contribution Low
Multi-device e2e specs (18-multi-device-sync.spec.ts) require a local Supabase stack to verify outside CI; no Docker daemon in some dev sandboxes Slower local iteration on sync code — Low
Mobile sidebar drawer (.sidebar overlay, transform: translateX()) always slides in from the left edge, even under dir="rtl" — translateX has no CSS logical-property equivalent Cosmetic-only RTL gap on mobile nav; inert until an RTL locale ships — Low
PWA manifest icons (SVG + 192/512 PNG) have no dedicated maskable purpose variant Android adaptive-icon masking may crop/pad the icon awkwardly on some launchers — Low

🎓 Development Notes

Testing Standards

  • Unit tests: Vitest (coverage thresholds: 50% lines/statements, 37% functions/branches — see npm run coverage)
  • E2E tests: Playwright, including a dedicated mobile-chrome and supabase project
  • i18n: Parity tests across EN, NL, FR, DE, ES locales — new locales must pass the same parity check and be checked against existing UI layout for overflow
  • Accessibility: axe-core in CI (WCAG 2.1 AA minimum; AAA contrast enforced via contrastCheck.ts)
  • Performance: Offline-first validation; sync latency benchmarks for > 100 rubrics

Last updated: 24 June 2026 — Phase 9 complete (9.1–9.4): RTL CSS logical-properties readiness, installable offline PWA (vite-plugin-pwa, prompt-based update, Supabase requests excluded from SW caching), a dyslexia-friendly reading mode toggle, and expanded essay exports (Markdown/DOCX/PDF, single/batch/combined-with-rubric). Phase 10 remains a themed proposal.

Clone this wiki locally