The discussion of American military innovation in the 1990s held out the prospect of substantial changes in the way that each of the military services would fight future wars. Capitalizing on advances in information technology, the military ideally would be able to decentralize operations (massing fires instead of forces), choose precise targets (minimizing the scale of destruction and especially collateral damage), and optimize strikes for political effect (integrating intelligence databases with tactical operations). If fully implemented, the tasks facing leaders and soldiers during both wartime and peacetime would change. Apostles of the revolution in military affairs (RMA) often downplayed the role of systems acquisition in the transformation effort. They argued that the main changes would be in organizational culture and training, affecting the way the military fights rather than what it fights with.1 They hoped that the trans formation project would not get bogged down in decades-long technology development efforts – efforts whose expense might make the revolution seem unobtainable. Moreover, innovation in commercial information technology (IT) visibly progressed at a blistering pace, and the price of civilian computing capacity plummeted every year. Technology seemed to be the easy part of the RMA. Some new equipment would of course be needed to add communications and information processing capability to existing platforms – to enable soldiers, sailors, and airmen to share the information that would allow them to fight differently. But at first glance, the core military platforms could stay the same, and the IT acquisition would be relatively easy. The services need only reach out and grab what they wanted from the incredible commercial market. Even if pure “off-the-shelf” buying rarely seemed reasonable, the whole point of IT seemed to be its adaptability, and few commercial businesses were immune to the whirlwind. So why should military buyers not also be able to get customized hardware and software for their needs with little effort? The acquisition side of the transformation effort would draw from a new set of suppliers with expertise in IT, whether established commercial suppliers or new start-up firms.2 These firms rather than the traditional defense industry would do the bulk of the relatively small development and manufacturing task for the RMA. But this relatively limited effort would not take full advantage of the RMA: different platforms, built from the ground up to depend on networks for their full

functionality, would be smaller, faster, simpler, and cheaper. If a platform could reach out through the network to draw on other assets for part of its capability – for example, if a sensor did not itself need to take a shot but could complete the “kill chain” by communicating with a separate “shooter” platform – then the platform would need to pack fewer systems into its confined space, and its form, fit, and function could be optimized for a single purpose.3