Like a Good Neighbor

The idea is smart, widely beneficial, fiscally responsible and—now that government technology officials have put it to work—obvious in a forehead-slapping, why-didn't-I-think-of-that way.

Instead of writing 12 different software programs to manage the inmate populations of county correctional facilities, Massachusetts will roll out a single data repository and program, with individual workflow, to each jurisdiction. The program will reduce development costs, ensure the sharing of appropriate inmate data, and focus more technical eyes on identifying and fixing bugs.

But the idea doesn't stop there. Peter Quinn, the commonwealth's CIO, believes that with slight variations, correctional facilities across the nation could take advantage of the same program. Quinn's idea: Share the program by putting the code in a Web data repository accessible only to government officials.

Massachusetts' taxpayers have picked up the development bill for this program, but improvements to the code made by other users might eventually help the commonwealth more efficiently manage its inmates. And other states might even share their own programs.

"Why not take the fruits of our labors and share them, when we’re licensed to do so?" Quinn asks.

It's a question that many public sector CIOs are asking. A handful of city and state technology officials joined Massachusetts in December 2003 to create the Government Open Code Collaborative (GOCC), an effort that promises to change how governments select and implement technology. The group's Web site and code-sharing data repository went live in June 2004.

"This puts the whole build-versus-buy decision into a brand new context," says W. Val Oveson, CIO of Utah, whose governor's office was a charter member. "We could save money and find software that's more relevant and useful. If it's already working for someone else, why build it from scratch?"

The GOCC is willing to share many useful applications, but legal, technical and cultural obstacles stand in the way. "This is uncharted territory, and some people are afraid to go here," says co-founder Jim Willis, director of eGovernment and information technology for Rhode Island's Secretary of State's office.

Common Bonds

Proponents say a public sector code-sharing cooperative makes sense because so many opportunities exist for collaboration. Inmate management, welfare, motor vehicle registration, taxation and a host of other service areas cut across state and local government boundaries. But CIOs often scratch and claw for funding to buy or build one-off systems to support these efforts.

"From our perspective, we are naive enough to think that because we do the same things and face the same problems, we should be able to share software," Quinn says. "We see opportunities for working together to incur lower expenses and enjoy the intellectual participation of a much larger community."

A similar effort in the federal sector provided inspiration for the GOCC. A U.S. General Services Administration Web site, www.core.gov, provides a central location for sharing software components among federal agencies.

On a smaller scale than Core.gov and the GOCC are some state efforts, including one pioneered by Utah. Seven states in the O-TRACK consortium use Utah’s inmate management system, a program similar to the one developed in Massachusetts. But this consortium exists only for a single application, not for the wide variety of applications and code that GOCC members want to share.

In Ohio, the Northeastern Ohio Technology Committee (NEOTC) brought together more than a dozen counties to share technology and expertise. (See “Bucks With More Bang,” State Tech, Fall 2004.) Four of these counties report that NEOTC collaboration has produced almost $2 million in cost savings or cost avoidance.

“These efforts are an attempt to use the open source model to become more efficient,” says Charles Garry, senior program director at the IT research firm META Group in Stamford, Conn. “Why reinvent the wheel, especially when a lot of governments do essentially the same things?”

Shared Applications

GOCC members point to several benefits of collaboration. Government entities that have large IT organizations, such as Newport News, Va., and Rhode Island’s Secretary of State’s office, anticipate that new features and bug fixes will result from code sharing.

Andy Stein, director of information technology for Newport News, helped develop a vote-tallying system that was uploaded to GOCC’s data repository. “If the software is useful and stable, others will use it, help us fix any problems and contribute their own enhancements,” Stein says.

In Rhode Island, the Secretary of State’s office contributed its “Town Crier” program, a scheduling package that posts online the times and minutes of government meetings. Willis of Rhode Island’s Secretary of State’s office looks forward to getting feedback from users of the shared software.

“We want as many people as possible offering comments,” he says. The group also plans to share a Web syndication program that aggregates government news about individual subjects, even when the information comes from separate departments.

Smaller cities and towns, such as Gloucester, Mass., don’t expect to be able to contribute a lot of code. “I’ll be very happy to let all the larger municipalities and states do the legwork on that front, while we keep something of a backseat. That’s my secret plan,” quips Mike Wells, director of information services. More seriously, Wells believes that small municipalities like Gloucester can fill the role of the expert user, providing guidance for application improvements.

Wells thinks it’s unlikely that GOCC’s open source programs will ever replace the large back-office financial or asset management applications municipalities buy from commercial vendors. “Probably no one at the municipal level is going to sit down and write their own financial package,” he says. “But it would be marvelous if a small open source application plugged into one of our larger commercial packages.” Wells also sees value in downloading “clever little fixes like Windows batch scripts that people are sharing—the kinds of things we might otherwise have to buy off the Internet.”

Possible Roadblocks

If collaboration is such a good idea, why hasn’t anything on the scale of the GOCC sprung up already? GOCC members cite several reasons.

Legal concerns top the list. GOCC members say the organization’s success hinges on letter-perfect adherence to software licensing contracts. “We are defenders to the utmost of everybody’s intellectual property,” Massachusetts’ Quinn says.

Accordingly, all code copied to the central repository must either carry a public open source license or be code that a public entity owns and legally can share. The rules sound straightforward, but lingering confusion makes compliance tricky. (See “By the Book ” above.)

States and localities also must run vastly different combinations of hardware and software, making for incompatible IT infrastructures. “Although on the surface we are building very similar applications for public safety, public utilities, tax systems and budget management, we developed our systems independently over the years,” Newport News’ Stein says. “Some of us are using COBOL applications, some are using flat files, some are using relational databases. It’s difficult to share an application from one environment to another.”

GOCC members believe that the larger open source movement is showing them a path around these technical incompatibilities. Members plan to develop “target-neutral” applications that can run unaltered in Windows, Unix, Linux and other common operating environments. “Open source code is the common denominator,” Stein says.

Others worry about potential service and support headaches. “One of the fears that I have heard expressed is that somebody will take our code from the repository and then expect us to help implement it, and that’s unacceptable,” says Utah’s Oveson. To address this issue, code contributors plan to specify that programs downloaded from the GOCC repository are “as is,” and public sector users or their commercial consultants are responsible for all support issues.

GOCC’s pending “as is” warning demonstrates that, far from excluding commercial software developers, code-sharing consortiums might open new revenue streams for them. “This offers a new model for making money for firms that specialize in adapting software,” Oveson says. Maintenance revenues are more profitable than licensing, with some large software vendors reporting profit margins of more than 80 percent for their maintenance operations, META Group’s Garry says.

A final roadblock to consortia success is skepticism bred by lack of knowledge. Some CIOs don’t always understand what open source software is and how it can help their organizations, Newport News’ Stein says.

“Until cities agree that we’re more alike than different in our needs, I don’t think [code sharing] will get much traction,” he concedes. “But once we start sharing valuable applications through the repository, and states and localities start downloading and using this software, I think they’ll start getting comfortable with how stable the technology is.”