It was the kind of detail that pharmaceutical executives at $20 billion companies don't typically bother pointing out. But Jacky Vonderscher, vice president and head of drug development at the Novartis Institutes for Biomedical Research in Cambridge, MA, paused while guiding a visitor through the company's spanking new research labs to sing the praises ofa hallway. True, it was open, airy, and exceptionally commodious, but all the same, it was a hallway, running through the institute's oncology and infectious-disease research laboratories. No one was likely to discover a new blockbuster drug in this hallway, any more than a scientist was likely to think up a cure for cancer while dreamily staring out the tall windows lining nearby labs.

"You see, it's not just laid out as a square," Vonderscher explained over his shoulder as he hurried along the corridor. Indeed, the hallway sliced through the labs at a provocative angle, revealing office geometry according to Picasso, not Mondrian. That architectural detail is just one small part of the multibillion-dollar attempt Novartis is making to provoke innovation by cutting across departments and disciplines, bringing people together in odd juxtapositions, and knocking down the walls between academic and entrepreneurial interests. In choosing Cambridge as the site of a new $4 billion operation, which will be the company's worldwide research headquarters, "We are trying to think of everything that affects the dynamics of the scientists and their interactions with each other," Vonderscher explained.

Like many other pharmaceutical companies, Novartis is in the midst of trying to revitalize the drug discovery process. Throughout the industry, that process has long followed a standard-and perilous-progression: identify a biological target like an enzyme or a gene that appears related to a disease; fling hundreds of thousands of compounds at the target in the hopes that several of them will interact with it; study the toxicity, absorption, and other physiological properties of the most promising molecules in animals; and if all still looks encouraging, test one of the compounds in humans. The entire marathon can easily take a decade and cost hundreds of millions of dollars. And actually getting across the finish line is particularly brutal: the failure of many potential drugs, for either safety or efficacy reasons, often does not become apparent until large-scale clinical trials involving hundreds of patients, which is why only about one in ten compounds that enter human testing succeeds and becomes a drug.