Council seeks software strategy

A Defense Department council this month will begin shaping a strategy to improve DOD's management of software programs in command and control, information management and weapons programs.

The Software Management Initiative, more than a year old, has already had an impact in a number of areas, including acquisition reform. However, this year the Software Management Initiative Council hopes to offer specific proposals and funding requests to address key management issues, such as educational programs for engineering personnel.

The initiative is being championed by Cynthia Rand, principal director for information management in the Office of the Assistant Secretary of Defense for Command, Control, Communications and Intelligence (ASD/C3I).

DOD has invested between $20 billion and $42 billion in software programs and shows no sign of slacking off, Rand said. "Hopefully we can make some changes before we get to the situation where we've got to do something `now' because it's so bad," she said.

Emmett Paige Jr., ASD/C3I, does not believe DOD has any more problems with software management than commercial industry, but it is an issue nonetheless.

"Software systems development and life-cycle maintenance continues to be one of the most costly and difficult tasks that we have to cope with," said Paige, who co-chairs the Software Management Initiative's executive board.

The Defense Science Board provided the impetus for the Software Management Initiative in 1994 with a study that spelled out a number of recommendations for addressing software management and acquisition.

The recommendations included the use of commercial best practices, expanded education requirements for DOD personnel, expanded use of software system architecture, commercial off-the-shelf (COTS) products and metrics. The board recommended specific actions on policy, education, software reuse and other software practices, and acquisition.

The council has spent the last 15 months combing through the Defense Science Board's recommendations, prioritizing some and deferring others. Six process action teams to address specific areas were established.

For example, one group is focusing on issues associated with DOD's increased use of COTS products, including security, integration and acquisition.

Proof-of-concept pilots will be a necessary ingredient in these efforts "so four years from now we don't find we developed a bigger, different problem," Rand said.

DOD needs to establish a baseline by which it can measure the success or failure of programs or management measures that spring from the initiative, she said. In the beginning, at least, such metrics will provide the council with a much-needed marketing tool—a way to demonstrate where improved management processes can pay off.

The council has endorsed a model called Practical Software Measurement, developed by the Joint Logistics Systems Engineering Group, for understanding the impact of software management improvements.

This model looks at how a management practice becomes part of the acquisition environment—adopted in terms of policy and use—and its measurable impact in terms of specific programs.

The council also recognizes education as a top priority.

"If we are in an environment where we want to rely more on the judgment of the program managers and [special program offices] and reduce the amount of processes they have to go through, needless to say additional training will be required," Rand said.

The group already has begun working on developing a software-oriented curriculum for personnel involved in software-intensive acquisition. Now the goal is to improve education programs for engineers.

This project involves working closely with industry to identify "core competencies" for software engineers and identifying DOD and public education programs available to DOD personnel. Rand also hopes to see more widespread adoption in DOD of the Capability Maturity Model (CMM), a technique for assessing a development organization's software processes, which was developed by Carnegie Mellon University's Software Engineering Institute (SEI).

DOD also sponsored the development of a similar model for assessing the processes of acquisition organizations; that model has already been used at a number of sites. That tool "is an important aspect of improving our management processes and reducing our risk," Rand said.

The biggest obstacle is the elusive nature of software itself. "Success in software is, unfortunately, something that is not talked about," Rand said. More often, she pointed out, people notice it "only when the hardware doesn't function properly."

The Software Management Initiative's executive board is co-chaired by Paige and Noel Longuemare, principal deputy undersecretary for acquisition and technology.

Board members include Anita Jones, DOD's director of research and engineering; Colleen Preston, deputy undersecretary of Defense for acquisition reform; and service acquisition executives from the Air Force, the Army and the Navy.

The council will have accomplished a lot just to bring out specific proposals to improve software management, said Scott Reed, manager of defense programs at SEI.

This initiative is only one of several that have been attempted in the last decade, but past efforts tended to fall short of actually generating "products," Reed said.

The next critical step is funding. In speeches this winter, Rand has described the initiative as "struggling for identity" in DOD.

Financial backing for specific proposals will give the council the solid footing it needs, she said.