AdaIC News Summer 1997


Letter from the AJPO

| Moving to a software-engineering process |
| The infrastructure | The lifecycle | The AJPO |


I am very excited and encouraged by the recent decisions made by the Assistant Secretary of Defense for Command, Control, Communications & Intelligence (C3I), Mr. Emmett Paige, Jr., regarding the recommendations of the National Research Council (NRC).

I specifically and enthusiastically support the decision to drop the Ada Mandate for all systems. Ada 95 is a superior software-engineering enabling technology. Ada 95 is extremely competitive when language choices are made based on technical merit and/or lifecycle costs.

I am fortunate to work for an extremely supportive chain of command that gives me the freedom to speak my mind frankly. Ada 95 makes engineering sense, and its use should not have to be mandated, regardless of the domain in which a system exists. Even the most helpful policy cannot effectively substitute for hard engineering analysis.

An Ada mandate is not needed if programming-language decisions are made as part of a rational software-engineering process. Language decisions for government software projects need to be made on a lifecycle basis to protect the interests of the taxpayers. The compelling DOD interest is to use Ada where it makes sense as an engineering solution.

Moving to a software-engineering process

There is legitimate concern about the implementation of the software-engineering review process recommendations of the NRC report. Given the difficulties associated with a discrete requirement such as the Ada Mandate, how can a more general software-engineering requirement be better implemented?

Right now, DOD decision makers have only just begun the process of deciding how specifically to respond to the NRC report. Still, we can see that a well-thought-out software-engineering process can provide objective focus for DOD software programs. Education and training investments can provide the means for DOD personnel to implement successful software-engineering practices. In this context, Ada 95 will continue to prosper.

We cannot effectively manage technology if we do not understand it. When we focus on management, we get better management efficiency. When we focus on process, we get improved processes. But if we want better software, it is best to be focused on software engineering. Program management and process improvement are non-trivial elements critical to the success of any program. But without software engineering, management and process cannot produce reliable, high-assurance software.

The infrastructure

An important conclusion of the NRC report is that the existing Ada infrastructure provides the United States with a significant competitive advantage in warfighting software applications. A viable infrastructure requires high-caliber, properly trained personnel, as well as tools and environments. It is very difficult to build this infrastructure in the government sector alone. Industry and academia have critical roles to play in maintaining and expanding the Ada infrastructure. Tri-Ada is one of the most aptly named conferences ever. Every year people from government, industry, and academia come together to a conference where their intersecting technical interests are addressed.

We know there exists at least fifty million lines of code in warfighting software that is written in Ada. This is an extremely conservative, verifiable figure. Failure to maintain an adequate Ada support infrastructure has serious military implications. A key observation of the NRC report is that it would be very difficult to re-establish the existing Ada infrastructure if it were lost. The warfighting ramifications are very serious.

Consider an unexpected deployment to a theater of operations with an extreme climate. Combat operations in such a theater reveal software shortcomings in targeting systems. If that software is not maintainable, we will face a potential war stopper.

The lifecycle

For those not intimately familiar with the warfighting missions of the Defense Department, it is easy to simplistically compare warfighting software with commercial software. However, this is comparing apples to oranges: Military software and commercial software are simply not the same. It almost certainly does not make business sense to use all the extra safety features provided by an Ada compiler for a commercial product that will be on the shelf less than 18 months.

The business case that applies to defense systems is the lifecycle argument. Although sometimes difficult to rigorously quantify, the software-maintenance advantages of Ada are not seriously disputed. With software-maintenance costs running 70%-90% of software lifecycle costs, it is clear that Ada usually wins any lifecycle cost comparisons.

The AJPO

The Ada Joint Program Office does not make policy, it implements policy. We have a very important role to play in the execution of the policies that are developed to implement the recommendations of the NRC.

In my view, the most important contribution we can make is in the education and training of our human resources. In the Defense Department, we have great people doing some very tough jobs with very limited resources. We can never go wrong investing in people. Education and training are force multipliers that will produce a sustained high return on investment.

John A. Hamilton, Jr.
Lieutenant Colonel, U.S. Army
Chief, Ada Joint Program Office


Previous Page Contents Next Page