With a stroke of his pen, Richard D. De Lauer has created a booming new software industry. DeLauer is the Undersecretary of Defense for Research and Engineering, and last June he signed off on a memo ordering that certain Department of Defense software be written in Ada, a new computer language. The order paves the way for Ada to become the only high-level language in defense software. Not surprisingly, it has heated up the battle for the Ada software market, which has drawn in companies as big as IBM Corp. and as small as $2-million-a-year TeleSoft of San Diego.
The decision -- effective as of January 1 -- has been in the works since the early 1970s when defense officials began to worry about the cacophony of computer languages in use throughout the Pentagon and its service branches. They figured that a uniform, all-purpose, departmentwide language would ultimately save taxpayers billions of dollars. So they began the search for an appropriate one, eventually coming up with Ada.
Ada is, in fact, the latest in the line of major programming languages that includes Fortran and Pascal. (The name comes from Lady Augusta Ada Byron, who was the daughter of the English poet Lord Byron, and who is widely regarded as the first computer programmer.) Like other languages, it is designed as a formal coding system using English-like phrases to specify the instructions for a computer to follow -- functioning between the level of application software (what the computer user deals with) and machine code (the binary language that the computer understands). But unlike the other languages, Ada is capable of serving the full range of military applications, from processing the Army's vast payroll to navigating submarines.
Because Ada is so comprehensive and abstract, it is also very difficult to use. For that reason some observers doubt that it will ever gain wide acceptance. "We just have to look at the record of failure with Jovial [a previous candidate for a uniform language] to know what the fate of Ada will be," says Charles Lecht, chairman of Lecht Sciences Inc., a New York City think tank, and a vocal critic of Ada. "I think it's a klutz."
From a business standpoint, however, it probably doesn't matter much whether or not Ada is a good idea. The fact is that the Pentagon has an annual software budget of more than $10 billion, including more than $3 billion for embedded software now covered by the Ada rule. If all that software has to be written in Ada, one thing is certain: A lot of money will be spent on Ada-related products and services in the years ahead, even if Ada itself turns out to be a failure.
Indeed, the DoD has already spent some $20 million on development of Ada, and that is only the start. As the language comes into use, more software will be needed to make it functional. Application software is expected to begin emerging late next year. Right now, the emphasis is on compilers, the programs that translate the Ada instructions into the computer's grammar of ones and zeroes. It is possible to write Ada software for any computer that has an Ada compiler.
One company active in this area is TeleSoft, which was formed in 1981 through the merger of two smaller software companies. Headed by chairman and chief executive officer Kenneth Bowles and vice-president Craig L. Maudlin, the company has already sold over 350 compilers to commercial and military customers. It also markets other types of Ada support programs.
Intermetrics Inc., a $32-million-a-year Cambridge, Mass., software engineering firm that does 60% of its work for the DoD, is working on a compiler as well as other Ada software for the Air Force, while $36-million SofTech Inc., a Waltham, Mass.-based software company is working on a compiler and software development tools for the Army.
Of course, all this effort is going to create a crying need for Ada-trained computer programmers. SofTech is developing a training course for the Army, and Computer Thought Corp., a Plano, Tex., company that is less than two years old, is offering computer-based courses in Ada.
Then, too, there are the nonmilitary markets to consider. The Pentagon's backing gives Ada more than a fighting chance to emerge as one of the dominant high-level languages of the late 1980s in the commercial software market. If it does, the demand for Ada-related products and services will be even greater.
Before that happens, however, Ada must establish itself throughout the military -- no easy task. The services are fiercely independent, after all, and they have resisted previous attempts to institute uniform languages. So, while business will certainly be brisk for the next few years, Ada may eventually -- like its predecessors -- meet its Waterloo.