2025-09-25 Frontend Working Group Meeting Notes: react-query key structure, frontend-base next steps
Date, time, location
Date:Sep 25, 2025 at 15:00 UTC (timezone converter)
Location: https://meet.google.com/wxe-myxy-uei
Discussion topic(s)
React-query queryKey factories and structure
Frontend-base next steps:
Reviewing/merging Gradebook, Account, Profile, and Discussions initial conversions (Adolfo Brandes , Jacobo, Javier)
Keeping tabs on frontend-app-instruct progress (Diana Villalvazo Salas)
Is react-query required, or just nice to have? (Javier)
Design tokens (Adolfo Brandes and Brian Smith)
Rebase of existing conversions on master, after Design Tokens support lands. Includes getting all tests to work. (Adolfo Brandes, Jacobo, Javier)
Lazy loading (Adolfo Brandes)
tutor-mfe support (Adolfo Brandes)
rsbuild vs vite discovery (looking for volunteers)
🎥Recording
Video: https://drive.google.com/file/d/1KR1viWd-reaXKNNpa-xj-qIjVlO5nFiB/view?usp=drive_link
Chat: https://drive.google.com/file/d/1bwSLr5W9RO47vCTKsKJlNQrl0cDT2jvE/view?usp=drive_link
Transcript: https://drive.google.com/file/d/1GQiLe3SPIZC308tDC11_5mnz5Yp9CJOMhsAOFpPYzgQ/view?usp=drive_link
Participants
Adolfo Brandes, Brian Smith, Diana Villalvazo Salas, Jacobo Dominguez
🤖 Summary
1. Query Key Factory Proposal
Context: Adolfo proposed improvements to how React Query keys are managed in the admin console MFE.
Key Points:
Suggested using query key factories instead of hard-coded strings.
Recommended structuring keys from most generic to most specific, starting with the app ID (e.g.,
oropanetics.frontend.app.learner.dashboard).Helps prevent conflicts across MFEs, improves debugging clarity, and ensures cache consistency.
Discussion:
Brian and Diana supported the idea.
Concern raised about app ID uniqueness if multiple developers use generic names (e.g., “checkout”).
Solution: use reverse DNS naming (e.g.,
com.opencraft.checkout) to enforce uniqueness.Adolfo suggested documenting this pattern later as an ADR, but starting with an implementation first.
2. Updates on MFE conversion
Current Work:
Jacobo and Javier are progressing on their MFEs.
Profile MFE: Jacobo took it over after Victor dropped out.
Gradebook and Account MFEs: PRs under review; Javier expects to push updates this week.
Timeline:
Several MFEs (including Dashboard) expected to be available experimentally for OMO in October.
Tight deadlines due to the upcoming Ulmo cut (October/November).
3. Design Tokens & Paragon Integration
Adolfo’s Work: Attempted to port commits from
frontend-platformandfrontend-buildto frontend base.Decisions:
Base Paragon styles don’t need to be loaded separately; only overrides will be managed.
This simplifies implementation while keeping consistency across MFEs.
Tooling:
Use the Paragon CLI (
replace-variablescommand) to update CSS variables.Goal: enable design token support across MFEs.
Next Steps:
Adolfo will draft a PR for frontend base and one converted MFE.
Brian to review and test with example themes.
Target completion: by next Thursday (Sept. 30) for initial implementation.
4. Build System Discussion (RSBuild vs Vite)
Comparison:
RSBuild: faster but less mature.
Vite: fast, widely adopted, more community support.
Consensus:
Favor Vite for long-term maintainability.
Need to verify compatibility with MConfig.jsx and plugin loading.
Adolfo and Brian agreed more evaluation is needed before committing.
5. Tutor Plugin & Frontend Base Strategy
Proposal:
Create a separate Tutor MFE plugin for frontend base instead of overloading the existing Tutor MFE.
Pattern would mirror the Paragon plugin: bundles hosted as extra static files via Caddy.
Benefits:
Cleaner separation.
Easier configuration via Django settings.
Supports flipping between MFEs without heavy rebuilds.
Closing Notes
Meeting ended on a positive note, with consensus on several technical strategies.
Emphasis on tight timelines (Ulmo release by late October/November).
Team agreed on a clear path forward for query keys, Paragon integration, and frontend base plugin structure.