January 1997: my first HL7 working group meeting
It was a dark and cold New England January in 1997 when I came down to my first HL7 working group meeting in Tampa Bay, Florida.
SGML
I was invited to talk about Standard Generalized Markup Language – SGML – an object of veneration of a small cult that would give the world both HTML and XML, the engines of the World Wide Web and what came to be called ecommerce. To put this in context, the Web was just over 3 years old and XML was still a suckling infant.
It had been announced the previous month, in Boston, simultaneously with a talk on “SGML in Healthcare” by me and clinical co-conspirators, John Spinosa, Dan Essin, and Tom Lincoln. We finished before the XML folks and caught the final few minutes of what we knew meant a dramatic change in electronic information publishing and processing.
Prior to XML, the electronic text world was torn between the advocates of HTML (it’s simple, you can use it!) and SGML (it’s powerful, you can move mountains).
For the HL7 Tampa talk, billed as a tutorial, I recruited two SGHappyML luminaries – Tim Bray and John McFadden – both Canadian, both showed up with winter colds. Tim was one of the prime movers behind XML; John later a reluctant convert. John had written the OmniMark language for markup conversion and he could actually make SGML work. Tim wrote one of the first programs to index the Web.
We were greeted with perhaps more than a healthy dose of skepticism. It came in the flavor of “anything you can do (with < and >), we can do better (with Objects)” and I got my first exposure to “you come from outside, you can’t possibly understand our world”. To the betterment of the industry, cooler heads, or perhaps heads whose feet were well grounded in the wide world of implementation, prevailed.
As I recall, the tipping point toward a not-invented-here syntax for HL7 came on the heels of Wes Rishel’s admonition that failure to do so would look foolish in years to come.
Fresh thinking is the path forward
I believe that we must constantly be revitalized by fresh thinking. There will always be a tension between solutions designed for unique versus general requirements just as there is enmity between the Perfect and the Good. Almost 20 years later, as XML is taking a back seat to JSON and Version 2 interfaces make way for RESTful APIs, the culture of innovation is stronger than ever.
Dedication and drive
Leaving my first HL7 working group meeting, I lingered awhile in the sunshine and toured the winter home of Thomas Edison, a man who knew how to get it wrong as many times as it took to get it right. I bought a reproduction 1893 bulb and installed it outside my house where it has shed light through 20 Vermont winters. It is a symbol to me of the dedication and drive within HL7 and that we keep going until we get it completely right. (And if that one ever burns out, now they come in LEDs!)
