Or will the current crop of software--and the underlying technologies that enable it--be dismissed as faulty, defective pieces of flotsam: mechanisms that are hopelessly flawed, tediously error prone, and riddled with innumerable bugs that make them cheaper to replace than to fix?

An Epidemic of Failure or Business as Usual?

Consider the following study conducted by the National Institute of Standards and Technology (NIST). NIST surveyed 10 CAD/CAM/CAE and product data management (PDM) software vendors, and 179 users of these products in the automotive and aerospace industries. It discovered that 60% of users experienced significant software errors in the previous year. Of those who reported errors, the users indicated that there was an average of 40 major and 70 minor software bugs per year in the tools they were using.

That doesn't sound too bad, does it? After all, even Microsoft discovers a bug in its code upon occasion. ;-)

But according to NIST, the total cost impact of these software errors on these manufacturing sectors alone was estimated to be $1.8 billion per year.

NIST conducted a second survey of the Financial Services sector (4 software vendors and 98 software users). This second survey focused on the development and use of Financial Electronic Data Interchange (FEDI) and clearinghouse software, as well as the software embedded in routers and switches that supports electronic data exchange (EDI). Two-thirds of the respondents--mostly major banks and credit unions--indicated that they had experienced major software errors in the previous year. Those respondents that did have major errors reported an average of 40 major and 49 minor software bugs per year in their FEDI or clearinghouse software systems. Approximately 16% of those bugs were attributed to router and switch problems, and 48% were attributed to transaction software problems. The source of the remaining 36% percent of errors was unknown.

The total cost impact on the financial services sector was estimated to be $3.3 billion.

In case after case, industry after industry, the use of buggy software is clearly having a detrimental effect upon the bottom lines of our companies.

59 Billion Dollars Worth of Software Bugs

The NIST study, entitled "The Economic Impacts: Inadequate Infrastructure Software Testing" concludes that the total cost of these software errors to the U.S. economy is an estimated $59.5 billion annually or about 0.6% of the gross domestic product. The focus of the report is directed at the need for a better software testing infrastructure to remove or drastically reduce the number of errors before software products are released into the economy.

But the identification of "the problem" doesn't really address the development of products within software industry. After all, if any software company is given an adequate budget, the easiest thing to do is to discover that something doesn't work properly. The more difficult challenge is to develop a system of technology that prevents software bugs from penetrating the development process to begin with. Object-oriented programming (OOP) was once touted as a means by which such software errors could be eliminated, but that has certainly not become the case. And today it seems that the dynamics and the economics of the software industry are making it hopelessly difficult to address the issue of software quality at all.

Constant Change Negates Quality Control

For instance, a typical software application today is invariably composed of thousands of discrete software modules and components that are interlaced and connected--and that often interact with secondary or tertiary functions embedded in the operating systems of the hardware. Middleware, operating systems routines, Web services, communications protocols, and myriad other technologies all interweave to theoretically provide the user with a transparent and highly functional experience that purports to automate or streamline work.

However, too often these software elements are in a constant state of change--through new software releases, updates, or patches--making the testing process of any particular component or package extremely complex. The failure of a single piece of code anywhere within the stream of commands that pulse through our systems may lead to false information, security leaks, or catastrophic failures that may not even be identified as "bugs" by the user. In other words, what is amazing about our systems today is not that there are so many failing pieces, but that anything works at all!

Yet clearly something is wrong with the scenario by which the industry is developing products. Consider that for any other industry--the automotive or the space industries as examples--a failure rate of 60% in any component would result in a national recall of the product. The software industry of today doesn't--and couldn't--operate in that kind of quality-focused environment. Why? Because the economics won't support it.

Rate of Change: 18 Months and Counting Down

The average turnover of products within the software industry is about 18 months. This means that every year and a half some major component of an information system will be upgraded, swapped out, superceded, or scrapped. This cycle of replacement is actually accelerating as competition within the industry progressively drives software and hardware vendors to the next level of technology. For instance, as companies move toward Internet-based Web services, the invisibility of component change--from a systems administration perspective--will make it impossible for any individual software vendor to thoroughly test and guarantee the quality of any particular product. In the realm of Web services, software updates and new releases will roll out to our companies' software infrastructure with the same alacrity as Internet viruses. Meanwhile, our ability to control or even test these changes will be erased or severely hampered.

Quality: The Industry Penalty

One might conclude that the only real remedy for this dilemma is for the industry itself to establish and maintain rigorous standards for the quality of software. But the computing industry has already debunked the concept of quality if it interferes with the competition for innovation.

Consider the experience of IBM: In 1987, the United States Congress passed Public Law 100-107 called The Malcolm Baldrige National Quality Improvement Act, establishing an award to recognize products of the highest quality developed within the United States economy. In 1990, IBM's Rochester Division won the award for the AS/400 computing system. Yet, instead of becoming a standard for excellence within the industry, the AS/400 was branded and penalized as a "legacy proprietary system."

IBM seems to have learned its lesson and has now turned the argument of standards and quality on its ear: Its marketing strategy is now to equate so-called "open-source" standards--standards designed to publicly identify the specifications by which technologies interact--with the concept of "quality standards."

International Standards Versus Standards of Quality

For instance, last January, while marketing both the e-Server hardware and the WebSphere middleware computing technologies, Bill Zeitler, Senior Vice President and Group Executive of IBM Server Group, claimed that no single organization can now control the momentum of e-business through proprietary hardware or software schemes. His message was aimed at hardware and software competitors who had built proprietary computing systems. In IBM's view, owning a proprietary technology meant that the quality of the product wasn't up to international standards.

Yet, the irony is that to control the quality of any product, a developer must take ownership of the process by which the product is created. This is, of course, Microsoft's argument as it pours R&D dollars into its proprietary .Net technologies. Furthermore, Microsoft says, there are no embedded standards of quality within the open-source movement. As a result, customers won't experience a decrease in the level of errors from these products: Instead, in all likelihood, each release of innovative software based upon open-source standards may actually see a net increase in customer software errors.

Software Development and the Loss of Innocence

Clearly the industry is headed into a new realm as it grapples with the competing dynamics of quality and competitive innovation. Long gone are the days when a small group of individual software developers might control the quality of every element of its products. Gone too are the days when the testing for the quality of a product can reveal all the flaws that a consumer might experience.

Perhaps when the history of this era of tool-making is written, it won't be the tools themselves that stand out as revolutionary. Instead, maybe it will be the processes by which we learn to control the transformation of technology itself. Until such a history is written, however, our companies seemed destined to swim in a sea of technological change, each of us learning to inhale and exhale new advancements in technology without drowning in the flood of technological errors and bugs. And until the software development process has been changed, it seems clear we'll continue to suffer from a surfeit of program bugs that rob us of productivity and steal our precious IT dollars.

Thomas M. Stockwell is the Editor in Chief of MC Press, LLC. He has written extensively about program development, project management, IT management, and IT consulting and has been a frequent contributor to many midrange periodicals. He has authored numerous white papers for iSeries solutions providers. His most recent consulting assignments have been as a Senior Industry Analyst working with IBM on the iSeries, on the mid-market, and specifically on WebSphere brand positioning. He welcomes your comments about this or other articles and can be reached at This email address is being protected from spambots. You need JavaScript enabled to view it..