TAG meeting at Ed-Fi Summit
Austin Hilton - Room 417B
Agenda (draft)
- Ed-Fi Community Intellectual Property Disclosure Policy
- Tier 1 Error Improvement (continued)
- See TAG Meeting 2019-09-11 meeting notes
- Course Transcript Data Model issue
- See TAG Meeting 2019-09-11 meeting notes and materials
- 2020 Technology Roadmap input
Participants
- Jeffco Public Schools – Rob Sukiennicki
- INsite - Rosh Dhanawade
- San Diego County Office of Education - J. Pablo Rodriguez
- Student1 - Geoff McElhanon
- Wisconsin DPI – John Raub
- Certica Solutions - Jim McKay
- Metro Nashville Public Schools - Lei Wei
- InnovateEDU - Marcos Alcozer
- Education Analytics - Andrew Rice
- Skyward Inc. - Josh Reimer
- Arizona Department of Education - Britto Augustine
- Infinite Campus - Jennifer Downey
- Ed-Fi Alliance, Eric Jansson
- Ed-Fi Alliance, Chris Moffatt
- Ed-Fi Alliance, Vinaya Mayya
- Ed-Fi Alliance, Sayee Srinivasan
Notes
2020 Roadmap
TAG feedback on 2020 roadmap is captured in the attached deck. The discussion was mostly around early themes, and did not go into great depth on any topics.
Tier 1 Error improvement
The discussion noted that not all errors are created equally, as some errors can result in many subsequent transaction failures. There are also common classes of errors, such as failed key unification. The TAG recommendation was to introduce error codes and "severity" levels for those errors likely to lead to many subsequent errors (aka critical "stop" errors) such as the failure of a school record to post.
A second recommendation was for vendors to put version numbers in HTTP headers, to assist in diagnosis of issues.
Course Transcript data issue
The TAG recommendation was to remove the Course reference and introduce a "Course Code" descriptor list. The TAG felt that a separate entity around "credit capture" risks confusion, does not capture the commonplace semantics of the domain, and displace the use of Course Transcript itself, resulting in a net loss of referentiality. However, the TAG agreed that some fix was needed.
A descriptor list was seen as a good compromise between referentiality and flexibility, and the developing concept of operational context could also permit mapping between contexts, allowing there to be state lists of codes and local lists of codes.