TAG Meeting 2019-09-11

Owned by Eric Jansson
Last updated: Sep 16, 2019

Agenda

  1. Housekeeping items (Eric)
    1. Summit update and headcount
    2. SIS and Assessment API RFCs (data model updates)
  2. Validations API SIG Update (Chris)
  3. Analytics Middle Tier – next version (Stephen Fuqua)
  4. Tier 1 Error Improvement (Jen)
    1. See attached doc
  5. Course Transcript Data Model issue (Eric)
    1. See attached doc

Participants

  1. Tom
  2. Jim McKay
  3. Marcoz Alcozer
  4. Juan Pablo Rodriguez
  5. Andrew Rice
  6. Chris Moffatt
  7. Geoff McElhanon
  8. Jen Downey
  9. Lee Barber
  10. Neal Schuh
  11. Rosh
  12. Stephen
  13. Vinaya
  14. Shannon
  15. Sayee
  16. Veronica
  17. Eric

Notes

Validations API - SIG results were summarized including the 2 options (summarized here: Report Out - Open Source Rules Engine and Validation API SIG). While the issue was not discussed in depth, those who spoke mentioned support for Option 1 - to work with an actual agency if going forward with implementation.

Analytics Middle Tier v2 was discussed, with major deliverables being compatibility with Suite 3 and support for descriptor mapping. This release was discussed in more detail. 

  • It was mentioned that this would not support extensions, but that possibly some support for standard extension sets backed by RFCs could be included (e.g. ED-FI RFC 18 - FINANCE API)
  • Maintenance concerns were raised - could be high overhead. It was pointed out that this overhead will be born individually by members of the community anyhow.
  • Performance and scaling were raised as a concern. It was noted that this likely needs to be on a downstream mirrored copy, and not on a system actively gathering data.
  • Another set of comments recommended to focus on highly reusable use cases.

Tier 1 errors. We reviewed the data shared by Infinite Campus. There remain some inconsistencies in errors and the proper action is to raise tickets and attach them to  ODS-3117 - Getting issue details... STATUS  Another action we discussed was to compile all errors and post a public list of errors and ways to address them, as part of the core development effort. Another possible action was to include more actionable error messages directly in the errors, but it was noted this is likely more difficulty as there may not be a clear place to intercept the errors in the current architecture. Issue will be revisited at the Summit.

Course Transcript model issues was not discussed - will also be discussed at the Summit meeting.


Materials