Silly season

What did you do this summer? I spent some of it with a cyborg cockroach.  Not to mention some computing devices constructed entirely of Lego, a Nobel Prize-winning cosmologist, the head of the European Human Brain Project, the director of the NASA Ames Research Center, a group of scientists searching for extrasolar life, another group trying to bring back woolly mammoths, the creator of a human-powered helicopter, a microbiologist who takes photos with bioluminescent bacteria, and the founders of Google.

If you think this sounds like a 21st-century version of Alice's Adventures in Wonderland then you wouldn't be far wrong, though in this case it was for real. I was attending Science Foo Camp ("Sci Foo" for short), a gathering of scientists, technologists and other thought-leaders held each summer at the Googleplex in California and co-hosted by yours truly.

As you may have gathered, Sci Foo isn't an ordinary scientific conference. What makes it special - apart from the remarkableness and diversity of the attendees - is the freewheeling, self-organising format. Foo Camps, like wikis, are an idea that do not work in theory, only in practice. And they were born in, of all places, the world of publishing.

In 2003, Tim O'Reilly and his colleagues at O'Reilly Media, a technology publisher and events organiser, were wondering how to make use of the excess space in their new offices in northern California following the dot-com bust.  The result was a weekend gathering of 200 technologists in which the schedule and topics for discussion were not determined in advance but rather by the participants themselves during the course of the event - an approach known as an "unconference". O'Reilly's Foo Camp (the "Foo" part stands for "Friends of O'Reilly") quickly became legendary in technology circles and a couple of years later a mutual friend of Tim's and mine suggested that O'Reilly and Nature (my employer at the time) join forces to hold a scienceequivalent. We quickly agreed and a subsequent email exchange with Eric Schmidt confirmed that Google were happy to host. Sci Foo was born.

At the time its success seemed far from certain. Several scientist friends opined that researchers tend to lack the gregarious nature of technologists, and have rather little to say to colleagues from outside their own narrow fields. But a short way into the first event in August 2006 we realised that our doubts had been ill-founded. On the morning of the first full day, a senior Nature editor inquired hopefully of me whether we intended to hold another one of these things. A few minutes later a well-known science fiction author was more forthright; he looked at me seriously and said, "If you guys don't do this again, you're insane", then walked off. In my mind, I punched the air.

Brain-busting

This year's Sci Foo was number nine in the series. As ever, topics were brain-bustingly diverse, from mathematically influenced art and the future of life (carbon-based, silicon-based or non-existent?) to scientific paradoxes and the reproducibility crisis in research. As night fell we relaxed by gazing at an amazingly clear view of Saturn, complete with rings and its moon Titan, and enjoyed a sip or two of beer brewed with yeast extracted from fossils.

Aside from an increased emphasis on demos (an innovation itself pioneered by early attendees), the format is virtually unchanged from that tentative first event eight years ago. So perhaps it's worth considering why it works. After all, the organisers of traditional conferences (and I am one of those too) put a great deal of effort into identifying relevant trends, crafting an agenda, and lining up just the right keynote speakers and panelists. Are they wasting their time?

Certainly, conference experience often reveals that the best discussions are the spontaneous ones that happen in hallways between sessions. But we're extremely uncomfortable basing a whole event on this notion, mainly because most of us would accept guaranteed mediocrity before risking the public humiliation of an outright failure - especially when it involves requiring hundreds of important people to travel long distances to attend.

And why take the risk in the first place - what business need does it serve? After all, Sci Foo itself generates no revenue. First, because hosting customers, collaborators and colleagues at an event that many come to describe as one of the very best of their lives does no harm to our standing in the communities we serve. Second, and perhaps less obviously, by allowing attendees to set the agenda we discover what is on their minds rather than simply hearing their answers to the questions we already think important. Many articles in Nature - and business ideas at Digital Science (my current employer) - have emerged from, or at least been greatly improved by, conversations at Sci Foo. Indeed, O'Reilly's whole "Maker" business arose from their observation that Foo Camp attendees who were previously working on software had taken to tinkering with physical technologies as well.

The lessons are clear: sometimes the most valuable thing you can do is to let go and allow amazing people to be themselves. And every now and then you should try something crazy. Who knows, it just might work.

Timo Hannay is managing director of Digital Science, a division of Macmillan Science and Education