Every PC user is familiar with the notorious "blue screen of death," the azure void that appears when Windows crashes. And even amateur geeks recognize the ubiquitous "Fatal Error!" warnings that pop up seconds before a program implodes, zapping into oblivion hours of spreadsheeting or a spirited game of Tomb Raider.

For the most part, these failures amount to little more than teeth-gnashing nuisances. But the consequences of buggy software are not always so benign, a lesson the folks at M.A. Mortenson Company learned the hard way. In 1993, the Minneapolis-based construction firm purchased a specialized program called Bid Analysis, manufactured by Timberline Software Corporation of Beaverton, Oregon. Months later, Mortenson used the software to prepare a successful bid for a Seattle hospital project. So far, so good -- until some old-fashioned pen-and-pencil number-crunching revealed that Bid Analysis had erroneously suggested a figure $1.95 million too low.

Mortenson blamed the miscalculation on a software glitch and alleged that Timberline had covered up the flaw. The company sued Timberline in Washington's Kings County Superior Court, alleging breach of warranty and seeking to recoup its $1.95 million. "Our position wasn't that you're guaranteeing the stuff to be perfect," says Mortenson attorney Bradley Powell, "but you've got to hold it to a negligence standard… It wouldn't have been that difficult for [Timberline] to send out a notice to its customers there -- 'Hey, there's a bug, don't do this, or do that.'"

But Timberline, with the help of a Microsoft-backed trade group called the Business Software Alliance (BSA), countered that the license attached to Bid Analysis, which disclaimed all liability, was ironclad. And the court agreed, tossing out the case. This past May, the Washington State Supreme Court upheld the dismissal, finding that it was not even necessary for Mortenson "to actually read the agreement in order to be bound by it." Everyone, the court concluded, should know that software comes "as is," like any used car or yard-sale tchotchke.

Mortenson v. Timberline was a landmark case in the software industry's campaign to secure a unique privilege: zero liability, no matter how negligent a program's design or implementation. The clever mandarins of the multibillion-dollar industry realize that a blanket exemption from civil liability will free them from the pricey headaches-tort lawyers, massive insurance-that plague other industries. Through the courts and state legislatures, they hope to achieve just that. And if they do, even as computer software runs more and more of the paraphernalia of our lives-from cars to smoke detectors-the cost will be measured in more than just money.

Once upon a time, the wordy contracts that accompany shrink-wrapped software were not considered inviolable. In 1991, for example, a federal court held in Step-Saver Data Systems v. Wyse Technology that a purchaser did not have to abide by onerous license terms if those terms were not revealed until after the product was purchased.

Of course, in new-economy time, 1991 was eons ago. In the interim, software titans developed into crafty political and legal operators-not to mention revered icons. Thus the BSA's involvement in Mortenson. And, perhaps more important, the industry's efforts on behalf of something called the Uniform Computer Information Transactions Act (UCITA), which would grant vendors sweeping authority to write one-sided contracts. Championed by the Digital Commerce Coalition-a lobby supported by such digital heavyweights as Microsoft, Compaq, and Intel-the law, which requires state-by-state approval, will be considered in all 50 states within the next five years.

Tellingly, the first two states to pass UCITA were Maryland and Virginia, both hell-bent on competing with the dueling Silicons, Valley and Alley. "People want to signal to high-tech businesses, 'We're a friendly environment, come here and stay here,'" say Jean Braucher, a University of Arizona law professor. "That's what actually happened in Maryland and Virginia. They played the two off each other-they said, 'If Maryland doesn't do it first, we'll move all the high-tech jobs to Virginia.'"

UCITA in a mind-numbingly complex-and, some grumble, sloppily written-document, and even the brightest legal minds can't agree on all its implications. But it would essentially buttress Mortenson, codifying the software industry's ability to write licenses as it sees fit. Users would have little recourse in the event of disaster; once they clicked on the "I Agree" button during the installation process, they would waive their right to substantive legal action. Perhaps most troubling of all, in a reversal of Step-Saver, software makers would not be obliged to disclose license terms until the product had already been purchased. Opponents sum up UCITA'S repercussions with a catchy aphorism: "Even if they broke it, you bought it."

Software giants say they need these protections because software is inherently buggy. But, while that proposition may sound feasible to the layman-how many Americans can write a single line of Perl of C++--it is a myth. Far from being inevitable, buggy software stems from software-industry business decisions. Intent on stocking Best Buy's shelves with newer and newer versions of their products, manufacturers often shortchange engineering "best practices": Programs regularly ship before they've been cleansed of myriad major flaws, from backdoor passwords to logic gaffes to viruses. Software firms also churn out dreck by loading products with bells and whistles or issuing "improved" versions that lure consumers into paying for pointless updates. As so-called bloatware is imbued with millions of additional lines of code, it invariably becomes buggier. In short, glitch-ridden software is prevalent because companies insist on including baroque digital flourishes, not because creating good coding is an impossible task.

The software industry's no-liability crusade would be bad enough if it only meant losing your data and screwing up your spreadsheets. But its implications are actually far worse. Off-the-shelf software in increasingly showing up in non-desktop devices. In the next few years, software will be embedded in everything from pagers to SUV engines. Programs responsible for life-or-death tasks have historically been subject to far more rigorous testing than their personal computer counterparts. But, as desktop software migrates into critical systems-say, hypothetically, a version of Windows that controls a Mack truck's brakes-the potential for catastrophe looms. Mortenson's $1.95 million debacle pales in comparison.

Another, less tangible, consequence of a liability-free software industry may be America's vaunted position as the global new-economy leader. Should Mortenson and the UCITA protect American engineering laziness, they would open the door for upstart European and Asian competitors. Save for a few select players, such as Germany's SAP, international software firms have yet to challenge American supremacy. But as expertise diffuses, particularly with the growing availability of open-source code, there is little doubt that other nations will gain.

The software industry counters that research and development would grind to a halt if Joe Q. Public could sue whenever Lotus Notes ate his data. But, as University of Miami law professor Michael Froomkin points out: "We do see technological progress in other industries where we have liability rules. They make new airplanes, right? And I think a Boeing 747 is much more complicated than most software. Clearly, making people liable for their negligence does not stop progress."

If fact, it hasn't even stopped innovation in the technology industry. Just ask Intel. In 1994, the hardware giant faced a wave of class-action suits over a defective Pentium chip. The prospect of ceaseless litigation forced the chip maker to institute a "no questions asked" return policy, and the bug has since been excised. Intel, needless to say, has continued to grow and remains the giant of the chip industry.

Mortneson attorney Bradley Powell was correct in conceding that perfection is impossible. But good software, like good 747s or good microchips, is not a pipe dream, regardless of what the industry's public relations machinery says. The companies simply need a good reason to make it. Unfortunately, if they continue to win the battles being fought in state legislatures and courtrooms across the country, they may not get one.