Banff, Alta.

The problems are so hard, they make your head hurt.

But the people working on them may hold the future of computing and information technology (IT) in their hands.

Three hundred of the world's top young IT students converged on the Banff Springs Hotel recently for the 32nd annual Association For Computing Machinery (ACM) international collegiate programming contest.

Tom Keenan, Business Edge
University of Alberta teammates Steven Soneff, Kevin Waugh and Devin Doucette, left to right, concentrate on solving a problem.

The contest represented the cream of the crop, drawing more than 6,700 teams representing 1,821 universities in 83 countries.

The students heard pep talks from event sponsor IBM about how technology will help us in ways ranging from instant language translation to speeding up mass transit systems. They took a junket to the Columbia Icefields, and in their honour, the Banff Springs buffet offered nachos and hot dogs instead of crab legs and caviar.

With the pleasantries over, the three-person teams set to work on some very gnarly problems (see web watch below, if you want to try them).

Some would say they were pretty contrived challenges, such as measuring a city's skyline or trapping a bad guy in underground tunnels. I suspect if you handed these problems to one of the Banff Springs' Japanese tour groups, you'd hear lots of "Wakarimasen" (I don't understand) and also "Naze" (why). Indeed, why should we care about trapping spies in imaginary tunnels?

"These scenarios - that you're a SuperSpy and there's a villain trapped in a labyrinth - disguise some very real problems," says Doug Heintzman, the director of strategy for IBM's Lotus division. "They're problems of grid theory; they're problems of series, a lot of mathematics and physics problems. What we're looking for is how people think through problems, how they divine the most efficient ways to find the answer to a particular problem."

Handling collaboration technology at IBM may be Heintzman's day job at IBM headquarters in suburban New York, but his passion clearly lies with encouraging the next generation of young people to love and excel at information technology.

A Canadian by birth and a Queens University alumnus, he was first involved with the contest in Vancouver in 2001, and for the last five years has been the event's chief sponsor and evangelist within IBM.

Heintzman says that IBM spends millions of dollars on the contest every year, and that no matter where in the world the event is held, most of the support comes out of IBM's lab in Toronto.

"We just believe it's very important to shine a bright light on this next generation of the elite problem-solvers in the world," he says. "Quite frankly, the whole society and economy in general has a bunch of very serious problems to face. We're going to need some really excellent problem-solvers."

IBM's Heintzman acknowledges a more parochial interest too, because his company is constantly trying to recruit the best young minds.

"We have laboratories and research facilities all around the world, and this is an extraordinarily dense and well-qualified recruiting graph," Heintzman says.

He adds IBM has already hired a large number of students from this year's regional programming contests. "In fact, we used up every single university hiring ticket that we had on hiring students from the regional championships."

Google and Microsoft chase these young geniuses too, so the world is kind of their oyster.

But for that all-important five hours, all the glory and future job prospects are eclipsed by the stark reality of trying to solve 11 devilishly hard problems in a giant room full of other wizards.

Kevin Waugh, from the University of Alberta, this year's host for the competition, has been involved before.

"If last year is any indication," he says, "I won't be nervous until five minutes before the contest."

Since only one team member can physically work on a computer at a time, the others design algorithms and write pseudo-code solutions in the team's notebook. Waugh says that the best teams don't really work together in terms of solving the problem, "but there's always resource allocation and stuff, and usually that all falls apart and everybody starts working together in the last hour anyways."

Of course, somebody has to create these problem sets and decide the right answers, and there's a team of question writers and judges for that.

Are the students ever smarter than the experts? Waugh recounts a story where "one of the Polish teams found like a three-line solution (to a contest problem) and the judges were all mad about it because they thought it was gonna be a few hundred lines and a lot of work."

IBM may have to wait a while to hire these guys, and yes the competition was overwhelmingly male, with only about two-per-cent female contestants.

That's something Heintzman says they're working hard to fix.

Despite his boyish looks, Waugh is already a master's student and is heading for a PhD somewhere, "probably in artificial intelligence and machine learning."

His teammate Steven Soneff is again going to work for Google this summer, then will head off to Stanford University for grad school in computer science and computer engineering.

Devin Doucette, the third U of A team member, isn't sure of his plans, but figures that a job in the Googleplex might be a lot of fun.

When the results were tallied, a Russian school, St. Petersburg University of IT, Mechanics and Optics, was the overall winner, followed by the Massachusetts Institute of Technology and another Russian institution, Izhevsk State Technical University.

The highest-scoring Canadian team was the University of Waterloo, a very respectable ninth place out of 100 teams.

Our friends from the University of Alberta solved three problems correctly, and got one wrong, for a 23rd-place finish, along with some great memories and a huge resumé boost.

Heintzman says the challenge for organizers is to squeeze enough problems of varying complexity into a five-hour period to differentiate between 100 of the world's best and brightest teams "who have already jumped over a whole bunch of hurdles to get here."

Yet, in Heintzman's view, the problems they see here are easy compared to what we all need to face in the future. "The real problems, you know, are how do we solve global warming or reduce the impact of a certain pandemic," he says.

But he's confident that the people in this room are going to be among those who tackle those really, really important problems using IT and their raw brainpower.

Web watch: icpc.baylor.edu/icpc/v2/ (Tom Keenan is a professor at the University of Calgary and an expert on technology and its social implications. He can be reached at keenan@businessedge.ca)