FHIR Connectathon Report

Sep 10, 2012

The first FHIR connectathon was held in Baltimore Saturday Sept 8th.

The following organisations participated in the connectathon:

In addition, there were a series of other observers through the day.

What happened?

Well, we sat around a table, and we started trying to exchange information between the various clients and servers that were represented at the table. Most of the servers are in the cloud (and available for you to do your own testing if you want), but it was still worth having people around the table, to create a focus, and to foster group-think around the problems that we were considering.

In keeping with the early stage of FHIR in it’s overall lifecycle, we regularly broke out into discussion about how the specification should work - particularly when we discovered differences in our implementation, or functionality that the clients needed that wasn’t described in the specification. We made a list of suggested improvements for the specification.

Towards the end of the day, we started to focus more on actually exchanging resources, and then we finished with a short demonstration of each of the clients at work.

Outcomes

These are the important outcomes from the connectathon for me:

  • This was the best fun that I’ve ever had at an HL7 meeting - I really enjoyed the day
  • We had a mix of greenfields software and adaptations of existing software, along with a wide variety of technologies from old to new
  • There was about an even mix of XML and JSON usage
  • Everyone who was present was able to exchange content with at least one other participant, and in most cases, several
  • The servers were much more complete and advanced than the clients, which are generally more proof-of-concept - this makes sense given where FHIR is. Our challenge for the next few months is to create domain content for the next connectathon
  • Ewout has developed a compliance testing application for servers, which he used throughout the day to drive closer conformance - this is a great tool that will really help implementation
  • The testing uncovered lots of bugs and surprises in the software, and we gathered a list of about 20 issue that will drive us to produce a better specification - with more clarity and easier to make it work
  • This is a really great way to do specification development - it gives us a real sense of what is important to actually fix, and it’s worth doing this again - the participants thought so as well
  • FHIR really does work - it really is easier.

Links

Acknowledgements

I’d specially like to thanks Mike Henderson who organised the connectathon, and provided refreshments at his own cost - and kept us focused on connecting to each other instead of getting distracted on arguments about XML vs JSON and REST vs the world etc - way to go Mike!

Also, I’d like to thank Chuck J for supporting the connectathon and coming by several times to see how it was going. In addition, thanks to everyone else who came by to check it out.

Finally, I’d like to thank everyone who came and demonstrated working software. It’s a big enough commitment merely coming to an HL7 meeting, let alone spending days or weeks actually developing real software that works: FHIR is fast to develop for - just like the name says - but it’s still a lot of work to do any software, so kudos to everyone who came.

This post brought to you by the insomniacs at HL7 who can’t sleep because of jetlag.