Anomalies Abound

by Lindsey Vereen

An old joke asks, "Who invented cottage cheese? And how did they know when they were through inventing it?" Software is like cottage cheese in the sense that it's hard to tell when it's done. The same is true of almost any sort of intellectual content. Fiction writer Flannery O'Connor was never satisfied with anything she wrote and was only published when her editor pried her manuscripts from her hands. Mike Nichols spent a year editing Catch-22, morphing the film from an epic with a cast of thousands into a metaphorical piece whose cast of thousands wound up on the cutting room floor. John Fogerty, of CCR fame, has been described as a perfectionist who drives the production cost of his albums up because of his meticulousness.

It's hard to tell when software is done, because it's hard to do. Eighty percent of a system's complexity is in software, and 80% of the development effort is software. With the increasing importance of software as the source of product differentiation, it seems curious that software companies are often so small compared to the hardware companies they support. While there are obvious exceptions to this rule of thumb, it is clearly evident among purveyors of embedded software development tools. Compare the size of the companies that build the tools with the size of the semiconductor companies they support. Tool vendors are often so minuscule, in contrast to the giants, such as Intel, Motorola, and TI; they are like the pilot fish that hover around sharks, waiting for morsels of food to come their way.

Software companies have often relied on the success of a single product for their survival. The classic example is Micropro International, whose WordStar program dominated the word processing market, and along with Visicalc and Lotus 1-2-3, created the PC market. Wherever you look, software is what translates electronics into a successful application.

Despite the value of software, most semiconductor companies have figured out that the hardware business is way more lucrative than the software business. For some reason they'd rather sell hundreds of thousands of chips than a few software licenses. Small companies have one distinct advantage over large ones: they can be very fast on their feet. That kind of flexibility can be especially valuable to an ISV during times such as these when innovation abounds. Small companies have the freedom to be innovative and responsive. Not only can they bring new tools to market quickly, but they can bring new types of tools as well to accommodate more complex design problems. As we're developing our 1997 Buyer's Guide issue, we're finding it a challenge to invent enough categories to accommodate all of the tools we want to include.

Speaking of challenges, we're looking for a new technical editor. Nicholas Cravotta is being pried away from Embedded Systems Programming by the publisher of a startup Miller Freeman publication. Finding someone conversant with embedded system development who can write about that subject in a logical, coherent way is not easy. If you live in the San Francisco Bay Area and have an inclination to explore the wacky world of publishing, drop me a line. You can help us invent each issue-and tell us when we're through inventing it.