Evolution Meets the Real World
William Toti
June 8, 2000
The impact of Moore’s Law is commonly misunderstood by system developers.

In 1968, Gordon Moore, then chief executive officer of Intel, correctly predicted that computer hardware, specifically processing capability, would continue to advance at a pace of about one generation every 18 months.

This remarkable prediction, which has come to be known as “Moore s Law,” has proved incredibly accurate. The law also has fueled a number of disparate movements in seemingly unrelated disciplines, such as in the arena of defense acquisition reform.

As the current defense reform argument goes, if information technology (IT) hardware evolves one generation every 18 months, how can we possibly allow ourselves to take 10 years or more to develop and field a weapon system that is heavily dependent on that IT?

That point is fair enough.

Certainly, Pentagon planners could do a much better job of engineering IT systems, interoperability and networks up front into weapon systems. The establishment of a DoD chief information officer with statutory purview over the architecture and acquisition of all DoD IT systems probably would help a great deal in establishing standards and protocols for data transfer and networks.

And the notion of creating a DoD-wide Global Information Grid that enhance s the ability to share information of all types to relevant players from weapon and targeting data, to payroll and business process information – is probably long overdue.

Once appropriate grid policies and processes are implemented, this initiative will undoubtedly go a long way toward improving weapon system effectiveness, helping to make the Navy’s developing concept of network-centric warfare a reality for all of DoD.

But while Moore’s Law may indeed have been useful in predicting the evolution of computer processors, using this concept as justification for shortening the defense acquisition timeline across the board is misguided.

This is true because the current acquisition reform debate is founded on a false notion, a perception that the development of information technology hardware drives the train when it comes to acquisition timelines.

Said differently, IT is generally regarded as the Holy Grail when it comes to developing future military systems. This belief has saturated conventional wisdom in the virtual world, but in the real world is a notion that is patently false.

In the virtual world, IT’s hardware evolutionary cycle may indeed be on the order of one generation every 18 months. But the same cannot be said for airframe technology, chemical propellant technology, munitions, radar, armor, wheels, target characterization, structures, imagery, hull, navigation, propulsion, seeing-through-the-trees, moving without noise, tracking rapidly moving targets, and other technologies.

These technologies evolve over a period of years, if not decades. What sense, then, does it make.to tailor the entire acquisition construct around one small component of a weapon’s overall effectiveness?

This means that as long as the defense of the United States has to be done in the real world instead of the virtual world, and as long as we properly continue to refuse to purchase yesterday s technology, then the evolutionary time line of computer hardware will only have an incidental effect on the overall defense procurement time line.

Further, the 18-month evolutionary cycle is only true for IT hardware. Software advances are nowhere near as rapid.

In the virtual world, we were promised advances in artificial intelligence technology a generation ago that still haven t appeared.

In the real world, you’re lucky if you can get high-quality, verified and validated code out of a software designer of a complex system in less than five years.

So why the rush to restructure the acquisition system? Its worth pointing out that the arcane, byzantine acquisition process we now enjoy is what produced the mightiest military in history, by an order of magnitude. Something must be working right.

The United States was able to do this because its· processes have traditionally been focused on real-world advances, not virtual advances. That is, the traditional U.S. acquisition system has been tailored to account for the tyranny of the real world.

The tyranny of the real world renders the tyranny of Moore’s Law an interesting but irrelevant factoid. Moore’s Law means that as you design your weapon system, you don’t decide what computer hardware to install into the package until after all the other hardware is designed, all the code is written, verified and validated, and other, more mundane features of the weapon systems are finalized. That is, insert computer hardware only in the final step.

At worst, trying to bend the acquisition process around one small piece of hardware the computer processor may result in breaking something that, all things considered, has worked rather well.

Moore’s Law is not without value. We should continue to teach it in college courses and speak of it in reverent tones in our Temple of Renewed Faith (Pentagon IT circles). We should evolve the use of technology so that we continue to advance toward that beckoning, ethereal goal of knowledge superiority.

But when it comes to the defense of the nation, we should restrain our zeal and limit ourselves to using only what works.