Make GPRA requirements high-level activity

Remember the Government Performance and Results Act of 1993?

It is supposed to be one of the drivers of improved program management in the federal
government, but we don't hear much about it these days. Either programs are already being
managed effectively, other things have become more important or no one is doing anything
particularly dramatic.

The first possibility would have been news. The second possibility, that there's
something more compelling, might mean agencies are investigating--and, we hope
avoiding--the impact of the year 2000 on software. If not, that leaves the third
possibility: Agencies are aiming only low-level activity at GPRA requirements.

But agencies have the means available to increase the quality of program performance
and save money.

In principle, the purpose of GPRA was to fundamentally change the way federal agencies
go about their work--with an emphasis on fundamentals. GPRA requires agencies to define
their missions clearly, set realistic goals, measure performance and report on their
accomplishments.

GPRA requirements are not something agency managers can tack on after the fact.
Performance measurement begins with setting program objectives and having measurable
performance metrics. Too many agencies have units that provide program evaluation, but
their tools are applied only after the program generates results.

What is needed is a lifecycle activity that begins with defining functional
requirements and stays with the program through final deployment. Most of the players in
the federal government know the activity as good old independent verification and
validation, or IV&V.

Unfortunately, few use it. That's too bad, because you can perform a credible IV&V
as part of a program and not get tangled with another bureaucratic oversight layer.

The Army's reborn Sustaining Base Information Services program has integrated project
teams that include test and evaluation units. They run tests using resources of the
program itself. The teams are charged with managing application development while assuring
the efficiency of existing ones. Although such processes cost money, the investment is
paid back as programs improve.

In some cases, agencies face mandated cost reductions. Validation is even more critical
to ensure efficient use of budget dollars. For SBIS, having validation teams has resulted
in fewer management reviews with more work getting done.

The Defense Department tends to give IV&V more attention in assuring quality
development and deployment scheduling of programs than do civilian agencies.

But the vulnerability to buggy software development and missed implementation schedules
affects nearly all agencies. Properly defined IV&V can enable programmers to meet
deadlines, stay within budget and otherwise perform as required.

NASA is expanding IV&V at Ames Research Center in Mountain View, Calif., to cover
software analysis and development research for the space shuttle and the international
space station programs.

A wide range of benefits result from IV&V implementation:

To obtain the highest level of objectivity, agencies have generally turned to
contractors for IV&V. More recently, agencies have discovered that they can break
ambitious programs into discrete pieces, then assign a separate contractor to perform
IV&V on each task in turn. The Immigration and Naturalization Service took this tack
for its Service Technology Alliance Resources System contract.

Agencies can use IV&V to uncover their year 2000 vulnerabilities while there's
still a bit of time left.

You can write follow-on program modifications that specifically call for the removal of
date code vulnerabilities. Fixes for 2000 problems take the form of several approaches,
all which can be accomplished through an IV&V program:

While none of these approaches is new, they can all be accommodated and managed through
IV&V.

Agencies that have already implemented IV&V programs cite some of their lessons
learned.

First, objectivity pays off. This is important because for most agencies, software
development doesn't have a good track record, and independent assessment helps identify
and correct problems.

Second, having a teaming and collaborating relationship with the IV&V contractor is
more successful than laboring under the old-style, arm's-length adversarial relations
between government and contractor.

Third, agencies experienced in using IV&V programs report that the money is well
spent, because they can demonstrate cost recovery.

There are no guarantees. After all, contractor IV&V and in-house development teams
need to do a good job. How do we know they will? For contractors, past performance may be
a good indicator. But each government program has different requirements and therefore
different solutions.

In a growing market, new players will line up, many without experience. Much of the
IV&V work has been through sole-source contracts held by companies with experience.
Proceed cautiously with unknown vendors.

In agencies where program managers have implemented IV&V, GPRA is well served, 2000
problems are minimized and, best of all, program performance is improved.

Robert Deller is president of Market Access International Inc., an information
technology market research, sales and support company in Chevy Chase, Md. His e-mail
address is bdeller@markess.com.