The alt-culture festival Burning Man has a motto – “No spectators” – meaning that anyone who attends is expected to actively participate, and the same motto applied to the first Ed-Fi Technical Congress that took place in April in Austin, Texas. Although there was considerably less body paint and fewer pyrotechnics than you’ll find at the desert festival (in full disclosure: my request for a smoke machine for the opening session was declined by management), the attendees who gathered in Austin will bear witness that much of the activity of the event was spent in the weeds of the Ed-Fi data standard and technology, debating designs, surfacing points where needs differ, and evaluating options for moving forward.

Beyond all the focus on technical details, the Technical Congress did have a clear theme: after two years of working to make standardized data exchange over REST APIs a reality, we as a community need to evolve our standards and technology. As adoption has grown from the first API usage in the states of Tennessee and Arizona, so have the use cases and needs. And while we did a number of things right, we also made a few mistakes that we need to correct, as well as ensure that we address new and emerging use cases.

So much of the discussion – both at the micro and macro level – was about how we as a community can evolve our standards and technology. In particular, how much change can we commit to, and on what timeline? From those discussions, here is what emerged as the major takeaways from the event:

  • Earlier release and longer cycles for Ed-Fi standards and technology are needed. Breaking changes are an issue, and the prospect of supporting multiple versions of APIs is a concern to the ecosystem. Participants recognized that ecosystem change has a long runway, and that vendors need time to “catch up” with a stable release.
  • Customization of data resources by SEAs remains a concern for Student Information Systems. Although ecosystem diversity in state use cases is not new, near real-time exchange of data via APIs seems to exacerbate the effect of multiple extension models.
  • Temporal data – particularly updating temporal data via standardized interfaces – requires more focus and input. Data exchange defined by “as of” dates is a practice today, and there was concern over availability of the data and of surfacing undue complexity in an API rather than via current manual systems.

These are not easy issues, and the community may not agree on proposed directions, or on how these questions will be implemented as technical standards. But the backbone of a healthy community is open dialogue, and the Technical Congress was an important milestone for the community in terms of tackling the tough problems at the root of interoperability.

The presentation materials from the sessions have been posted here.

A big thanks to everyone who gave their time and made the first Ed-Fi Technical Congress a huge success.