Back at the end of May, I posted 'Looking Ahead to SemTech;' an attempt to distill some advice for those visiting the event for the first time. We followed up a few days later, convening the team from the Semantic Link in front of a live audience of over 500 SemTechers; the vast majority of whom were new. And then, as the jet lag subsided, the Semantic Link returned to the more familiar environment of a Skype call to look back over the week that over 1,200 conference attendees had just experienced.
And now, as the jet lag from another trip to California recedes, I'm left with one over-riding impression of this year's SemTech; a growing connection to reality. In a week of conversation with strangers brought together by a shared need to wrest the last dregs from the coffee pot, I found a group of individuals for whom — increasingly — arguments about turtles and owls were a bemusing irrelevance. Microdata, microformats, rich snippets, RDF/XML, JSON, SPARQL and the rest, it appeared, were mere tools in their toolbox rather than the rallying cries around which competing Crusades might coalesce. Oh, the pragmatism.
There were, of course, still plenty of people who cared deeply about the technical and philosophical minutiae, and there was a lot for them on the programme. It felt to me, though, as if this group was (slightly?) in the minority, overtaken by a cohort from business and government with real problems to solve, real money to spend, and a real desire to fit semantic technologies into the set of tools at their disposal. People like this have been to SemTech before, of course, but (and this may be an unfair characterisation) previously they tended to already be semantic technology enthusiasts; they were searching for internal projects in which they could marry their private enthusiasm for the approach to the needs of their employer. This year's crowd seemed different. They already had the real projects, and they were evaluating semantic solutions alongside other ways of tackling the problem. This pragmatism, this desire to select the best tool for the job, was refreshing and extremely welcome.
This surfaced to a degree in discussion around the Google/Yahoo/Microsoft announcement of schema.org. The site appeared the week before SemTech, and was mentioned in various forms throughout the week. Reaction broke into several forms, encompassing everything from "They're not doing it right!" through "We're Doomed!", "It's Not Fair; they didn't consult with me!" to "It's not perfect, but if it gets more people familiar with publishing structured data on the web then great." Support for something like the final perspective, which is probably the one I personally prefer, seemed strong.
Reality also featured heavily in the keynotes. John O'Donovan of the Press Association, Dennis Wisnosky of the U.S. Department of Defense, and Bill Guinn of Amdocs all told stories of real deployments; of money saved, customers served more efficiently, projects delivered faster than previously possible. Library of Congress CIO, Laura Campbell, also keynoted on the final day, but I missed that one.
We need more stories like theirs, and we need more case studies of successful deployments where semantic technologies were fairly selected because they were right for the job, not because the budget holder was desperate to find a project where they could build an ontology in OWL or store data in a triple store.