Good Riddance, Win 98

By Peter Coffee  |  Posted 2003-12-15 Print this article Print

People in technology industries talk about exponential growth the way teenagers talk about sex. Everyone pretends to be having it, everyone wishes they knew how to get it and none of them fully understands the implications.

People in technology industries talk about exponential growth the way teenagers talk about sex. Everyone pretends to be having it, everyone wishes they knew how to get it and none of them fully understands the implications. The result, when either process does take place, is that bad decisions at the beginning are drastically magnified along the way.

What got me thinking about this—about exponential growth, I mean—was this weeks termination of all sales of several long-lived Microsoft products, including Windows 98. Whenever a platform product gets euthanized, I get angry mail from eWEEK readers who depend on third-party software or specialized devices that may be orphaned by the products demise. Overall, though, Im relieved.

Lets be sure were using the same word to mean the same thing. "Exponential" does not mean "fast." A quantity has exponential growth, over time, if the amount of growth from one moment to the next is in proportion to the amount itself.

If something grows at 8 percent per year, for example, then in year 0, it could grow from 100 to 108; in year 1, from 108 to 116.64. Over 10 years, simple growth of 8 units per year gets you from 100 to 180; exponential growth of 8 percent per year gets you from 100 to 216.

The difference is vastly greater when we deal with double digits. If we start at 100 and grow by 35 units per year, in 10 years were at 450; if we grow by 35 percent per year, in 10 years were at more than 2,000. But now we come to the implication that constantly seems to escape us, only to be painfully rediscovered again and again.

If we have 100 initial customers for a product, and 250 customers after three years on the market—a comfortable growth rate of just over 35 percent per year—we can hope to have more than 2,000 customers at the beginning of year 10, of whom 1,750 will come to us during year 3 through year 9. Seven-eighths of our customers, at that point, will be people that we attracted after three years of learning at the expense of our products early adopters. For how long, we must ask, are we obligated to keep on making the distinctive traditional mistakes that attracted the customers who got us off the ground?

To put it another way, whose needs should an IT vendor try harder to meet? The one in eight customers who liked what was first delivered, enough to tolerate its grave limits and weaknesses? Or the seven out of eight who required something better and who came to that vendor only when that became available? And who really do hope, moreover, that their IT sources will constantly do better—and who will otherwise move on to someone less encumbered by the past?

Loyalty to a products first users stems from laudable principles, but its not without costs. My favorite example is the "make" utility program, used by software developers to automate the rebuilding of applications with many components. The "make" program does the almost unthinkable: It treats the "tab" character as a distinctive symbol, instead of being merely one of many varieties of "white space." Complications ensue when other text-processing software ignores the distinction.

The explanation, admits the utilitys original author, is that "the pattern newline-tab worked ... and a few weeks later, I had a user population of about a dozen, and I didnt want to screw up my embedded base. The rest, sadly, is history."

Next page: A purer gene pool

Peter Coffee is Director of Platform Research at, where he serves as a liaison with the developer community to define the opportunity and clarify developers' technical requirements on the company's evolving Apex Platform. Peter previously spent 18 years with eWEEK (formerly PC Week), the national news magazine of enterprise technology practice, where he reviewed software development tools and methods and wrote regular columns on emerging technologies and professional community issues.Before he began writing full-time in 1989, Peter spent eleven years in technical and management positions at Exxon and The Aerospace Corporation, including management of the latter company's first desktop computing planning team and applied research in applications of artificial intelligence techniques. He holds an engineering degree from MIT and an MBA from Pepperdine University, he has held teaching appointments in computer science, business analytics and information systems management at Pepperdine, UCLA, and Chapman College.

Submit a Comment

Loading Comments...
Manage your Newsletters: Login   Register My Newsletters

Rocket Fuel