George Amana is a programmer/analyst for Tower Lawn and Garden, Inc. Tower is a distribution center for lawn and garden equipment in northern Louisiana. George just sat down to lunch in the company cafeteria. Pete Wilcox, a senior partner in the firm, joins him.

Pete Hey, George, you look pretty frustrated. What's the problem?

George I had to take over the sales information systems project that Judy left behind when she quit. It's total chaos. I was told that it was all but finished. Come to find out, she didn't finish several of the programs.

Pete But she did do a good job of specifying all of the program requirements. What's so tough about the programs? Judy always preached about the benefits of structured programming. In fact, she taught me how to do it. Don't tell me she doesn't practice what she preaches.

George No, her code is very well structured. And her documentation is adequate. It's just that the programs seem so poorly designed. Some of her subroutines are so long and complex that it's difficult to get a grasp on small enough pieces to test them for correctness. It seems like an all-or-nothing proposition. If I encounter a bug, I have to test large sections of code to zero in on the problem. Sometimes the bug turns out to be in an entirely different subroutine!

Pete Why didn't Judy break the system into smaller pieces?

George Oh, she did! The modules are evidence of that. But it almost seems like she generated the modules on the fly - as if to say, "Well, this piece of code is getting complex, I'd better put it in a module to finish it." She left me a rough draft of a structure chart, but I just don't understand the reasons she factored the system the way she did.

Pete That's the way I write programs, I start by trying to draw a flowchart on a single page - sort of the high-level flowchart. Then I factor the more complex processes into more detailed processes that I implement as modules. It sounds like that may be what Judy did.

George Maybe she did. But that strategy causes the lower-level modules to be very dependent on other routines. I frequently encounter bugs that get traced back to other, seemingly unrelated routines. I'm just getting further behind schedule. I may just have to rewrite the programs from scratch.

Pete Why don't you get some help? Barbara just finished her project. Maybe she can help you. You could divide up the work and get it done faster.

George Divide up the work? I don't see how. Judy's program specifications are just one big unorganized document. I'm not sure which file and report specifications to match up to which modules. For that matter, I'm not sure the programs themselves are fully documented.

Pete Well, I'm sorry, George. I don't know what to tell you.

a. What approach did Judy use to create the program modules?

b. How should modules in a program be conceived?

c. What effect does the program and module size have on testing?