By Michael Putzel
othing is more frustrating to the average computer user than software "bugs" that make a program misbehave in unpredictable ways, causing lost time and--worse--lost work.
Yet, crashes, lockups, "invalid page faults," "general protection faults" and technical failures beyond counting are taken for granted--even dismissed--by eminent developers and marketers as the necessary cost of winning the race to market in a feverishly competitive industry.
Stephen Manes, in a biting satire of the boastful speeches at the annual computer show called Comdex in Las Vegas, described the scene as it should be portrayed:
"On a huge screen, a montage of error messages dissolves into excerpts from baffling manuals and helpless help screens." Manes wrote in The New York Times. "To the beat of a busy signal, a Web page full of meaningless graphics crawls glacially down the screen. People sitting at computer screens slap their foreheads and pound keyboards frantically. A technical support representative wearing a telephone headset calmly asks, 'Have you tried reformatting your hard drive?'
"'Admit it: We tend to deliver products that are neither easy to use nor stunningly reliable,'" Manes's honest developer admits. "'So how have we managed to survive and prosper? Two words: fault tolerance. I'm not talking about fault-tolerant computers. I'm talking about fault-tolerant consumers.'"
In defense of the status quo, Steve Ballmer, executive vice president of Microsoft, said in a recent interview, "I'm not trying to say there's some excuse for bugs, but the reality is, you're always making a set of tradeoffs about the probability of problems--unknown problems--versus when you ship." Ballmer was explaining to the online news service news.com Microsoft's decision to declare Internet Explorer 4.0 a finished product and ship it in the fall of 1997 despite a stack of bug reports that clearly indicated the product was unstable.
Elsewhere in the technology industry, where lives may depend on the reliability of the hardware and software, there is no such tradeoff. I'm not suggesting that every piece of consumer software should undergo the debugging and documentation process of a program used aboard the space shuttle. But there must be a reordering of priorities to serve consumers who don't volunteer as beta testers when they buy new products. They have every right to expect their purchases to perform as advertised.
I installed the Internet Explorer 4.01 fix when it became available some months after the initial release Ballmer mentioned, and even the supposedly repaired version crashed the first four times I tried to use it. I don't mean to pick on Microsoft; the syndrome is broader and deeper than the products of the industry leader.
"The real problem with PCs today is their sheer unreliability," wrote Lawrence J. Magid in the Los Angeles Times. "Things go wrong at an alarming rate, and when they do there's often nowhere for a user to turn." Magid, a certified "techie" who nonetheless writes in plain English, said he hears from hundreds of frustrated PC users every month who are having "problems that can't be resolved by their software companies or hardware vendors."
I cling to the hope that as millions of nontechnical consumers pour into the personal technology market they will demand better quality. I thought Intel's public relations crisis over the flawed Pentium chip was a harbinger of new, more demanding standards.
Magid disagrees. "The sad fact is that things are getting worse, not better," he wrote. "I can't remember ever having trouble starting my Apple II or original IBM PC in the early '80s, before hard disks and sophisticated operating systems came in. The original Macintosh isn't as fast as today's Macs, but it was much more reliable. I'm not pining for the good old days--computers were less versatile and harder to use back then. But I am challenging the PC industry to come up with machines that are as reliable as our TV sets, telephones, refrigerators and other home appliances.
"Don't think you have to be a computer novice to be baffled by these problems. I know my way around PCs; I've even built a few. But I too get stuck now and then. I spend countless hours trying to debug hardware and software. And I lose productivity and sleep when things suddenly go wrong. Just ask my wife how many times I've been late to dinner or to bed because of some type of system crash."
Walter S. Mossberg, the noted writer of the Wall Street Journal's seminal Personal Technology column, blames what he calls "two cultural attitudes that make these stunts likelier than they ought to be."
"First, many people in high technology maintain a self--congratulatory belief that they are doing the hardest, most cutting--edge work in the world, so it's OK to sell products or services with known defects," said Mossberg. "Second, many people in the industry still labor under what I call the 'science--project mentality' of computing. This is the attitude that all this stuff is so new and experimental that the customers can be thought of as trial--and--error testers, instead of people expecting to buy finished products that work."
Certainly there was a compelling case for the latter point a decade or so ago, when personal computing was in its infancy and bright young kids were inventing programs overnight with all the energy and excitement of cyber pioneers. The so-called "early adopters" of technology learned the limitations of the products they worked with and frequently figured out their own solutions.
I recall traveling around the world with a tiny Radio Shack Model 100 laptop and a toolkit containing a pair of alligator clips that, with a healthy dose of trial and error, empowered me to break into almost any hotel telephone system in order to transmit my reports back home. There were frustrations, sure, but what great triumphs when you got it to work!
There was a relative handful of home computers in those days and a certain buyer-beware attitude. Today, there are tens of millions of consumers with PCs, and they just want their machines to work without their having to study an impenetrable manual or waste hours on hold or search vainly through a company's Internet Web site, only to find there is not yet an answer to the question that drove them there in the first place.
It's not too much to ask.