Tuesday, January 24, 2017

FHIR Connectathon has changed and it is good

I have been unable to attend HL7 WGM for a year, a problem that is now better.  This means that when I attended the FHIR Connectathon 14, prior to the HL7 Workgroup Meeting in San Antonio TX, I was shocked to experience the new FHIR Connectathon. This is good change on many levels. The others from the FHIR core-team have seen the changes over-time, so the transition is not as shocking as it was for me.

In past years, FHIR connectathon was more focused on people trying to use the core FHIR specification. This resulted in discussions to help understand the core concepts, core interaction models, core vocabulary, core data-types, and core Resources. This also resulted in corrections to the FHIR specification, so that future people would understand right from the specification. Other discussions were on toolkits, and how to make the toolkit better. There were discussions on how to implement a general purpose FHIR server. There were discussions on how to implement clients, starting from general concepts. The rumbling within the room was very focused on the FHIR core, and basic understanding.

This year, FHIR connectathon was more of a 'workgroup' meeting. More specifically, an "Agile" "Open-Workplace". There were a 16 distinct tracks. These are focus areas. All of them are broader concepts than one Resource, most are more of a 'workflow'.
Each of these were centered around one or more round tables. Where the table included various implementers interested in the topic, and a leader for the topic. What happens is very natural, they would discuss what the goal was, and adjust the goal organically as they progressed. They would try the basics that are described in the above links, but they would evolve based on the results. 

What really was a shocker is that Security and Privacy became a topic at most of these tables. 

Historically we have just pointed at SMART for security. SMART is a 'good enough' solution for the purposes of hackathons. SMART is not likely to be a good enough solution for production. This is a topic to be covered in much more detail in the coming months. SMART is now a project within HL7, so it will need some formal recognition, and framing. 

There was also interest in the basics of service discovery, patient data location, policy bridging, service response behavior, etc. All the kind of things that an operational environment will need to figure out. None of these were decided conclusively. BUT the very fact that they were talked about is evidence that FHIR Connectathon has changed... and it is good... 

FHIR has matured, it is growing up. It is not yet ready to fly away from the nest. But it is definitely no longer an infant.