A new information systems initiative quietly building up steam in the U.S. Department of Justice demonstrates that when public officials assume a more networked view of the world the result can pay dividends -- in this case, the promise of a much more effective justice system for all levels of government in the United States.

The seeds of this initiative began about a year ago when Paul Kendall -- general counsel to the Department of Justice's Office of Justice Programs (OJP) -- began to consider the future of law enforcement information systems and the billions of dollars being spent by federal, state and local agencies to build them. In his policy oversight role, Kendall recognized that information is the lifeblood of the criminal justice arena and few areas of government offer a higher potential payoff from the application of networked computing. It was also clear as he talked with colleagues across government that much of the horsepower in criminal justice technology had been squandered on "stovepipe" systems incapable of sharing data, and that policies and turf regarding information systems are often a roadblock to a more strategic vision of an integrated justice infrastructure. California got a painful dose of this when it recently discontinued development on a failing $350 million Statewide Automated Child Support System.

Kendall and his boss, Assistant Attorney General Laurie Robinson, saw this as a growing problem in light of the federal government's expanding grants -- now reaching hundreds of millions of dollars -- to state and local agencies for new information systems. Would the net result of these tax dollars be a future populated by hundreds of incompatible islands of technology, or was there an opportunity to build an information architecture that could seamlessly meet the needs of many criminal justice components, including law enforcement, corrections, probation, courts, prosecutors and public defenders? Kendall and Robinson decided to act, not by issuing a decree from Washington, but by assuming the role of catalyst and instigating a nationwide, "bottom up" discussion about the potential for a truly integrated justice enterprise.

OJP toured the country early this year, rounding up the "best and brightest" integrated justice initiatives from around the states. A great deal of support came from places like California, North Carolina and Texas, already struggling with their own plans for integrated justice networks.

In February and March, teams from eight states met with OJP in Washington at an informal summit to share ongoing work in their own states and to compare notes on how to best forge a workable justice enterprise. Participants at these meetings came from beyond traditional law enforcement roles and included professionals from across the justice arena, including judges, prosecutors and public defenders. Those involved recognized that broad representation is critical to an initiative of this nature as justice is a network of many distinct functions -- local police, county sheriffs, state law enforcement, correction departments, parole officers, prosecutors, public defenders and, finally, the courts -- which must effectively work together while maintaining the integrity of their own discrete functions.

OJP's initial meetings are showing positive results and, for the first time, a broad intergovernmental and interdisciplinary discussion is under way on what it will take to build a cooperative vision of a networked justice enterprise. The payoffs for achieving this goal are great. Clearly the biggest challenges Kendall, Robinson and their intergovernmental colleagues face will not be in the technical details but in how successfully they all navigate the human networks upon which a successful justice enterprise must ultimately rest. *

Letters to the Editor may be faxed to Dennis McKenna at 916/932-1470 or sent via e-mail to . Please list your telephone number for confirmation. Publication is solely at the discretion of the editors. "Government Technology" reserves the right to edit submissions for length.