Discussion XBlock Meeting Notes - 09-30-2015

Owned by Andy Armstrong (Deactivated)
Last updated: Oct 06, 2015

Attendees

Andy S (Unlicensed), ChristinaR (Deactivated), Steve Magoun (Deactivated), Xavier Antoviaque

Notes

  • We discussed what the T&L team is thinking for discussions
    • The proposed Discovery - Discussion Redesign epic
      • The intention is to keep this story light by building upon the /wiki/spaces/LEARNER/pages/19006021 design
      • The initial thinking was to update the existing xmodule without converting to an XBlock
      • Plan is to use the mobile team's new Discussion API
    • Redoing the discussion back-end
      • Replace the Ruby-based comment service with a more maintainable solution
      • Reevaluate the buy vs build options
      • Consider having a single discussion service that combines the comment service and Django comment client
    • Open question: does it make sense to do the UX redesign first?
      • Major benefits to user experience
      • Will it require a lot of rework once the new back-end is in place?
    • None of this work will be in Q2, but initial planning and architectural designs should start
  • For Solutions, managing their forked discussion XBlock is expensive
    • McKinsey solution cannot use XModules, so an XBlock was developed
    • The XBlock is a clone of the edx-platform code (primarily the client-side JavaScript)
    • Every time the platform changes, the code has to be merged into the fork
    • The preference would be to merge back the code so that McKinsey and edX are both using the same XBlock
    • This is a large project, so jimcaiN (Unlicensed) had suggested that a good first step would be to merge the code back as is
  • What are the options going forward
    1. Status quo
      • Solutions continues to manage the XBlock in their McKinsey fork
      • T&L builds the new redesign by updating the existing xmodule
    2. Merge the XBlock back to edx-platform
    3. T&L build the new redesigned discussion forums as an XBlock
      • Would be more work than just updating the existing xmodule
      • Major benefit: removing tech debt by building the new component following our own best practices
      • Supports a gradual rollout where most courses stay on the xmodule until the new xblock has proved itself
      • Would need to involve McKinsey at some level with the design and rollout of the new forums
  • Conclusion