TAG Meeting 2021-10-21 - Data Model Issues Raised by TAG

TAG Meeting 2021-10-21 - Data Model Issues Raised by TAG

Descriptors and Indicators 

DATASTD-1480 

DATASTD-1459 

Improvements to descriptor design: combining indicators and descriptors into a new core pattern.

  • Generally these failed to resonate and were seen as inviting some risks. There was no significant motivator for change.
  • Recommend abandon these ideas: too much change, too little known benefit.
  • TAG members reaffirmed that the Period pattern is awkward and not working.
  • Generally, TAG input was that simpler is better: just use a date to allow for simpler version. Forcing the invention of a date in some cases is not ideal but workable.

Likely outcome: abandon these ideas

On current descriptors:

  • Ed-Fi descriptors missing definition, short descriptions and long descriptions are all the same.
  • Would like system for how descriptors can be aligned to other standards. Makes it easier to adopt.
  • Operational context data could be integrated

Likely outcome: continue current Alliance practice of adding definition when editing specific sets if the semantics can be improved

Flexible Referentiality

Notes

  • A better approach from a “optionally-present” reference is to slim down the key and add an optional reference that – if supplied – is not “optionally present”
  • A “maybe it is there” approach could be particularly bad for LEAs – standards should be more principled to bring reliability to data exchanges

Likely outcome: abandon this approach

CourseTranscript, specifically

  • this means removal of Course from the key, and instead rely on CourseCode
  • Further, the Course Code should be local

Likely outcome: change CourseTranscript key to be a code and define it as local; possibly come up with a simpler system for state code transmission

Section Identity and Semantics

DATASTD-1596

  • Section should be more of a scheduling and not a course concept
  • In general, the approach recommended was to slim the key and add optional references to the problematic key components
  • Combine with the discussion on Flexible Referentiality the general advice can be seen as “lower referentiality requirements and ‘let the data flow’ easier”

Likely outcome: change section key to make required reference to CourseOffering optional, making it more of a scheduling concept