Repository navigation
Assignment 4
Release 1.3
Links to Submitted Files
- Updated use cases
- Updated sketch for reading karaoke UI
- Use case 5: AI-powered reading passage generation sketch
- UX research report
- Backend API documentation ADR
Progress and Contents Summary
In this release, our team has made significant progress in conducting our first UX study and defining our MVP scope. This submission includes:
- Updating our use cases with MVP specifications #38 #39
- Report of our UX research #40
- ADR for our use of Azure's API Management #36
Summaries of Issues Addressed
The following issues were addressed since the last release:
- DEV-3
- Description: Create initial backend service
- Resolution: Created Hello World Django app
- DEV-35
- Description: Add contributing guidelines
- Resolution: Contributing guidelines document created in Confluence, team requested for review.
- DEV-36
- Description: Implement BE API documentation
- Resolution: Setup automatic doc generation using drf-spectacular on our BE app.
- DEV-39
- Description: Research API management options
- Resolution: We've decided to use Azure's API Management.
- DCS-43
- Description: Update use cases
- Resolution: Updated the
product_research/use_cases.mdfile with our MVP scope
- DCS-38
- Description: Update UX research script and slides
- Resolution: Fixed a number of errors in our mockup slides and improved our script for performing UX studies
- DCS-40
- Description: Complete (remaining parts of) UX research report
- Resolution: Report completed and included in the team repo as a PDF
- DCS-41
- Description: Change wording in use cases paragraph
- Resolution: Fixed minor wording issue in speaking karaoke use case
- DCS-33
- Description: Write ADR for BE API documentation framework
- Resolution: Use drf-spectacular as our API documentation tool for the backend
Changes to the Roadmap, Architecture, or Use Cases
Roadmap Updates
We have integrated feedback from the teaching team's review of our use cases into our roadmap.
In accordance with the teaching team's recommendation to drop some use cases, we decided to drop the long reading and convobot use cases from our MVP. The epics associated with these two use cases have been reordered to after the end of the course.
Based on our team's capacity and the wow-factor delivered, we chose to keep the reading karaoke use case over long reading. To faciliate faster delivery and greater technical clarity for the work related to the karaoke feature, we've split the work into multiple epics. Some of these epics are scheduled for the future (i.e. after the course) because we will not include them in our MVP. Furthermore, some tickets have been created or changed within these epics to reflect our decided MVP scope.
Based on the teaching team's feedback, we have also split the AI-generated text feature into its own use case and corresponding epic. Lastly, we rearranged some tickets concerning the dashboard use case to reflect our decided MVP scope.