ADAPT and Learning to Get Along

20th Anniversary

ADAPT and Learning to Get Along

By Paul Schrimpf, Marketing and Communications Director

Fifteen years ago, the early concepts of what would become the ADAPT Standard version 1.0 released last year was a glint in the eye of a handful of agriculture companies. And how we got where we are today is yet another testament to the value of AgGateway and its active member volunteers.

It wasn’t easy. The early days of “precision agriculture” were alive with big breakthroughs, but also hyper-competitiveness across ag organizations, in particular machinery manufacturers and software companies.

Virtually all of the innovations being brought to market were riding in on proprietary software programs designed to operate effectively within a family of machines and software, limiting its value to farmers with mixed fleets. Could the industry come together and find a way to make interoperability at some level possible?

AgGateway had already begun to chip away at incompatibility in 2010 with a handful of companies coming together to form the PrecisionAg Council. Member volunteer Stuart Rhea was with FMIS company AgConnections at the time, an early arrival to the Council, and recalls being pulled into these early efforts.

“I got pulled into the SPADE project, which would be the first data interoperability project of its kind,” says Rhea. “SPADE went through three phases of development and generated the data requirements for ADAPT to get started in 2014.”

The two years that followed were full of both contention and collaboration as the industry worked through the challenges. “There were plenty of arguments and it got a little combative at times,” says Rhea. “We were going through the exercise of looking at each of the objects that were in the data model and figuring out what they ought to be, and what properties ought to be on them. It was a lot of work, but we got it where it needed to be.”

Rhea describes the process this way:

“The initial kickoff meeting for ADAPT was in mid 2014, again assembling the subject matter experts and data architects from the Council; but this time each company was encouraged to bring their internal data model. And most did. This is the miracle that is AgGateway – a place for fierce competitors to work together in a pre-competitive space for the good of the ag industry.

The resulting “superset” object model enabled every participant to express the data they wanted to exchange but there were a few problems. First, the object model was huge. It was simply too bulky and complex for developers to use. Second, it was overwhelmingly biased towards North America. For example, the Product object contained a place for EPA number, which means nothing to people using the model anywhere else. Over the course of about six months a solution was found. The model was split into two pieces, a lightweight, universal common object model that contained only the data that was needed by everyone, and an extensible system of tags that allowed additional data (like the EPA number) to be attached as needed.”

“One of the guiding principles of ADAPT was to make consuming the data as easy as possible. To that end, a plugin framework was developed that allowed each company to be responsible for mapping their proprietary data structures into the common object model. This plugin was then loaded at runtime and did all the heavy lifting.

The approach had the benefit of allowing each company to continue evolving its own proprietary formats while shielding the consumer from changes. It also gave each company the power to decide how they wanted to license the plugins use. Most of the participating companies used C# and .NET so that was the language selected to implement the model and plugin framework. Two community supported plugins were subsequently developed, one that translated ISOXML (ISO 11783-10) data and another that serialized to file the in-memory object model. Version 1.0.0 was released February 5, 2016.”

Rhea points to Dr. Joe Russo, then president of a company called ZedX, for setting up public demos of the proofs of concept that helped companies visualize and understand the work being done, ultimately leading to industry adoption. He also notes that John Deere’s investment of dollars and expertise was critical for moving the ADAPT Toolkit development forward to its ultimate release.

At that point, the machine data being collected was about helping to tune the performance of the machine. “The value of the data collected was really an afterthought for most companies until later. You’ve really seen a pivot from that now. Today, that attitude is completely different because they've realized that the competitive advantage the differentiation comes from what you can do with the data, not in holding the data itself captive. In my opinion, that is a pretty significant evolution.”

Fast forward to 2022, and the ADAPT Toolkit was ready to advance to a full global standard. Extensive use of the toolkit and a wide range of feedback both positive and negative drove the initiative. “ADAPT has been maintained and extended over the past eight years by a large, collaborative group of equipment manufacturers, farm management information system (FMIS) software companies, and universities,” says Rhea. “All of them recognized that the grower and other agricultural users have a critical need to use data from multiple sources in field operations decision-making.”

Feedback received from all over the world was compiled, and in mid-2022 the ADAPT Committee convened a new working group of subject matter experts from across the industry with the goal of removing any barriers to adoption and innovation. The result was the new ADAPT Standard version 1.0 last year.

“ADAPT version 1.0 features a further simplified common object model expressed as JSON schema, vector data expressed as Geoparquet, and raster data expressed as GeoTIFF – all of which enable development with a broad array of programming languages,” says Rhea.

The new standard is yet another example of AgGateway’s critical role in facilitating industry collaboration to drive connectivity and interoperability.