June 25, 1989

THIS YEAR’S PC Expo at the Javits Center in New York City essentially put the B in boring. Not that a show for volume buyers is necessarily boring by definition (although it is by me), but as one editor said, this show has never been known for products. Hardly anyone of note from the PC industry even attends the panel seminars, and those who do are speaking to half-full houses.

Downstairs from the Expo was a much smaller and more interesting show called National CASECON. Computer-Aided (or Assisted) Software Engineering, or CASE, is beginning to significantly affect software development and, if history repeats itself, it is likely to end up having a big effect on all computer users, not just software designers and programmers.

CASE is one of those terms that, like multimedia, means whatever the person who happens to be saying it wants it to mean. Barry Glasgow, vice president of Knowledge Based Systems for Shearson Lehman Hutton in New York, is president of a new organization called the Metropolitan New York CASE Users Group. He says CASE is trying to bring “more rigor” to the method and concept of constructing software.

CASE tools are not very sexy to people who are used to the wild, hands-on types of applications found in the PC and Macintosh world. But until CASE, Glasgow says, software engineers have been like shoemakers who don’t have time to patch their own leather — they’ve been spending their time increasing the productivity of the people they work for, and no time at all increasing their own.

“The last group of professionals to be automated was the software engineers,” he says. “Programming itself is an application. We needed a set of tools to do software engineering.”

CASE tools automate functions such as defining the requirements of a new program, then move on to functional and technical design, debugging, testing, quality assurance, implementation, operations and (very importantly) maintenance. Some CASE products even generate source code, the language of characters and text that make computers operate.

But in the future, it seems clear that CASE will take on a deeper meaning. As software shakes out into clearly segmented functions, and vendors (hopefully) start creating and adhering to standard specification for hardware and date formats, it seems inevitable that users will be able to pick and choose from the kinds of functions we need and “snap” them together into exactly the kind of program we want.

Glasgow likens the trend to the way people looked at telephones in the early 1900s. “In 1910, someone analyzed the penetration of the telephone into America and said that by 1930 there would be 90 million telephones and everyone would have to be an operator,” Glasgow says. “That’s actually happened. The process has been so automated that today we are the operators,” doing the dialing and disconnecting, assessing other networks, and even switching calls or forwarding them to another location.

Once upon a time, phone operators probably sneered at the thought the same way some software developers sneer at CASE. Software developers are opposed to anyone thinking that programmers are other than artists. That’s for good reason. There is a great deal of art to writing good programs — it’s very much like writing good fiction, only they use a more arcane language.

But proponents of CASE are often working with mainframe software — millions of lines of code built over many years, doing many mundane tasks that are anything but artistic and that need regular tune-ups, just like a car. Too many drabblers could turn a software tune-up of that size into a corporate nightmare.

Besides, there will always be art.

Productive Imagery Corp. of Berkeley announced a visual software development program for MS-DOS called ObjectVision at CASECON. President (and program author) Brain Sawyer, who also wrote VP Expert for Adams Osborne’s Paperback Software, says he designed the program so regular folks could use his visual, paint-type interface to make diagrams that are real, working programs which can link-up to standard databases and the whole bit.

“I wanted to do a couple of things,” says Sawyer. “First, I wanted to make it easier for people who don’t know how to program, and I also wanted to make hackers more productive.”