I found an easy solution to the resounding silence that sometimes happens, especially with uptight audiences, when you leave a nice big block of time for questions, close the monologue, then wait with rising anxiety for the hands to go up... umm... anyone? I now deliberately leave out a few fundamentals that everybody wants to know. By the time the floor is open for questions, hands shoot up for the FAQ, and snappy well-oiled answers get the room warmed up for the good stuff. Works even with the stuffiest crowd.

One piece of advice I'll add is to know as much as possible about the audience and the client's expectations. I once stumbled into an awkward situation with a keynote at a Fortune 500 international annual sales meeting (nice gig via my speakers bureau). Just as I was about to go on, the client walked over, introduced himself, and said, "oh, by the way, please avoid regionalisms and wordplay. OK, looks like you're on!" The audience was divided into blocs by country, with many wearing headphones getting live translations from workers in other rooms. Yikes. My usual engineering-humor patter, sprinkled with anecdotes, led to some awkward moments as different groups would laugh, frown, frantically take notes, or just stare blankly back at me. Took a while to adapt to that room!

Third, if at all possible, know more about the subject than anyone in the audience. Before my "claim to fame" that led to lots of gigs, I would occasionally accept conference-session invitations. More than once I would look out in the audience and see the people who created the technology I was talking about... which can be a bit intimidating. Not until I was actually speaking about my own thing did I fully relax. In situations where there is some overlap between the two extremes, I'm happy to engage domain experts in the audience.

Oh, and Berkun's advice to videotape yourself is golden - I did that early on and was horrified to see, when fast-forwarding, a cyclic physical behavior pattern that appeared with almost robotic regularity. Along with uncovering a couple of speech habits I didn't know I had, that tape was immensely helpful in debugging my presentation style.

Sorry, old words. Data Processing. Armonk, New York, is where IBM mainframes were made. And Amdahl was another mainframe competitor.

Showing my age, I guess.

Ages ago (60s or early 70s), a large aluminum company built a new HQ building (in Richmond, IIRC). They ran the numbers on computer-cooling vs building-heating, and made the computers an integral part of the equation (downscaling the heating plant accordingly). You see where this is going...

As the move approached, the DP guys saw an opportunity, and canceled their PO to Armonk... opting instead for an Amdahl, I believe. Winter came, and people started wearing coats at their desks. My friend who worked there reported that they were hastily building a kluge auxiliary heating plant with insulated ducts running across a parking lot.

Of course, the Docklands project doesn't sound like it's making any assumptions about the amount of waste heat, just doing something useful with it. But I hadn't thought of that paleo-computing tale in decades and had to pass it along.