Practical Programmer – Communications of the ACM

"We have met the enemy, and he is us." —the 1950s comic strip character Pogo

So you think there’s too much "fatware" in the software world? Software packages that have far more features and capabilities than the average user needs or wants?

Well, I do. I could have gone on happily, using my ancient WordStar word processor, for example, if the computer on which it ran hadn’t up and died on me. (It’s an interesting example: hardware wears out and breaks, but software never does. Witness all that Y2K software still running long after its programmers decided that two-digit years would last the lifetime of the program.)

Ah, but I digress. I want to talk about fatware. So back to my thoughts on word processing.

I’ve come to appreciate some of Microsoft Word’s new features and the newer breed of packages. But, to be honest, I don’t think I’d miss those features much if the world suddenly stopped, and I had to return to those days of WordStar yore.

So why do software vendors engage in the rampant featuritis that results in fatware? Why do we have the 800-pound gorilla products, instead of the lean—and perhaps mean—ones we wish we had?

Turns out there’s a good reason for all of this, one I haven’t seen discussed before in the computing press. It’s a reason rich in business tradition, and one with a striking analogy to other fields of human endeavor.

Consider the annual automobile model changeover. Each year, sometime around the fall in most cases, automobile manufacturers bring out a new model, one different enough from last year’s model to attempt to lure auto owners into "trading up." The underlying thought here is that the vendor is creating an excuse for the consumer to buy something new. Swoopy-edged styling? That sure obsoletes my old Scandinavian-lean model. Cup holders galore? How did I ever function without them? Oh, boy, it’s time to trade in the old dog. How could I go on without all these improvements?

Or consider the styling changes in clothing, especially women’s clothing. Each year, designers parade a set of models down runways all over the world, showing off clothing designs offering nothing much different from last year’s models except that, well, they are different. And each year, buyers flock to the shows, salons, and stores, eager to look up-to-date and not last-year dowdy.

Do you see the analogy? Commerce finds tricks to convince shoppers they need the new, to eschew the old. How else can an enterprise keep its sales curve ascending monotonically?

What can the software vendors offer as their bag of similar tricks? New and better features.

There was a time, of course, when some of these changes made sense. The annual auto model changeover dates back to that time—not so long ago—in automotive industry years, when advances in automotive engineering really did produce better cars, cars worth trading in. All-steel tops and hydraulic brakes in the 1930s. Automatic transmissions in the 1940s. Significantly better fuel economy in the 1970s. Dramatically increased reliability in the 1980s. Computerized componentry in the 1990s.

Commerce finds tricks to convince shoppers they need the new, to eschew the old. How else can an enterprise keep its sales curve ascending monotonically?

But between these dramatic improvements were long periods when style was just about the only motivation the auto industry sold. In fact, back about 50 years ago, the U.S. auto industry was resisting making any kind of changes except stylistic ones. That was the time when the book, The Insolent Chariots (John Keats, 1958), took the industry to task for playing games with the consumer, selling old wine in high-finned new bottles. It took high gas mileage and high-reliability Japanese cars to shake things up in the U.S., changing that pattern. U.S. cars could no longer afford to be "Insolent Chariots."

Now, let me ask a question: At the beginning of this column, I mentioned using WordStar. How did you feel about that? My guess is you had some mixed emotions. Part of you probably said, "Gee, that guy is really out of date. I’ll bet he also uses a 286-based PC and writes his software in Fortran or Cobol." Another part of you may have thought, "Gee, what a stubborn old coot he is." My guess is that few of you thought: "Gee, there’s a man of principles, one who sticks to what he believes in." Another guess is it probably didn’t occur to you at all that "there’s a way to show the software vendors we don’t like fatware."

We’re all conditioned. For the most part, we buy the latest clothing styles (never mind that last year’s styles are just as functional). We drive a fairly late-model car (never mind that most cars today run reliably well past 100,000 miles, and in fact many don’t even need any kind of engine tuning until that point). We buy the latest, greatest, multimeg computing products. And—for all of the same very human reasons—we buy fatware.

Is there a point to all of this? Is there a solution to the fatware dilemma? To be honest, my answer is "probably not." As long as we continue to choose style over substance in the rest of our lives, we’ll probably choose fatware over leanware when it comes to software.

Personally, I don’t believe software vendors are doing anything wrong in perpetuating this problem. For example, this is not the kind of thing that claims of antitrust violation can be built on. I don’t like fatware, as you already know. But the Microsofts and Oracles and Computer Associates of this world are simply following the rich traditions of such pioneers as General Motors and Gucci. They’ve figured out a way to keep selling new copies of old products, year after year, by taking advantage of this strange pattern of human behavior. They’ve simulated the style-based, annual-model, changeover-focused, retail model that’s been around for probably centuries.

"We have met the enemy," in this world of software fatware. And, as the comic strip character Pogo so profoundly said, "[it] is us."