Taking a Hard Look at Software

After more than 27 years in the information systems business, I never fail to be amazed at companies with mission-critical applications strung across a variety of platforms, all jerry-rigged together, which were never intended to communicate. As you indicate in ''Software Hell'' (Special Report, Dec. 6), the risk potential increases geometrically as the complexity of the communications and hardware configurations grow.

Far too many companies have accepted these risks for far too little gain. Many feel that the application of immature technology is necessary to maintain a competitive edge. But nothing alienates customers and internal users more than technology that does not work. Reliability and performance, once the gods of data processing, have given way to the personal computer and network server with highly volatile and largely unproven software involved.

In every era of change, particularly on the scale of the so-called Information Age, a period of time is needed to consolidate and digest the ramifications. Such a time appears to be upon us.

Geoff Wascher
Utica, Mich.


You give a much-needed public warning of a problem much lamented by those of us who make our living writing embedded software. While it is possible to design a simple product without design defects, some defects seem to be inevitable in any complex product. Unfortunately, business realities frequently exacerbate the problem.

Most software bugs reach the marketplace because a company knowingly ships a defective product in order to reduce the time to market. This is practical only because most software is sold under the Uniform Commercial Code, which allows a manufacturer to sell usable but defective merchandise at a low price without the usual implied warranty of merchantability.

With software being sold as defective merchandise, the only real protection that buyers have is the constraint offered by criminal laws, particularly those relating to fraud and bait-and-switch advertising. Removal of many of these protections will return us to the days of caveat emptor.

The first step in fixing Software Hell is to make fixing it a business necessity. Rather than weaken the law, strengthen it by forbidding the knowing manufacture of a defective product and forbidding disclaiming the normal implied warrantee of merchantability except for disposal sales at a loss. Is there anyone (other than an attorney) who can justify selling a product at a profit while claiming that the product is not fit to sell?

Wayne Albin
Snohomish, Wash.


''Will bugs eat up the U.S. lead in software?'' (Special Report, Dec. 6) nicely characterized the software quality problem and the fact that U.S. industry has been slow to recognize and address it. Unfortunately, it also implied that I was single-handedly responsible for the work to address this problem.

While I did initiate and lead the work to produce the first version of the capability maturity model, it was fully developed by a joint effort of the Software Engineering Institute (SEI), U.S. industry, and the Defense Dept. My more recent work on personal and team software process improvement also involves a team of SEI professionals and a growing number of industry and academic participants.

The U.S. needs more people willing to devote themselves to address this problem. It is thus important to recognize those who are already participating in this work and to encourage more to join.

Watts S. Humphrey, Fellow
Software Engineering Institute
Carnegie Mellon University
Pittsburgh


As the systems and product manager for IBM's OS/VS1 during the 1970s, I had the opportunity to work in Watts Humphrey's Intermediate Systems organization. At that time we were emphasizing zero defects in our operating system. We used many different techniques to test the programming code we received from development. We measured it against analytical programs used to determine the projected number of problems per thousand lines of code.

Because of the guidance and support we received from Watts, we were able to ship the highest quality operating system (at the time), as measured by customer acceptance and problem reporting.

It is interesting that Watts had to leave the U.S. in order to demonstrate to the software industry that quality is not only important, but problems can be detected and corrected prior to shipment.

Richard G. Weiss
Boynton Beach, Fla.


The author of ''Software that doesn't work'' (Editorials, Dec. 6) was obviously not around during the good old days. The IBM software of the 1960s was every bit as bad as the Microsoft software of the 1990s. Every month a reel of tape containing thousands of program temporary fixes (PTFs) arrived from IBM. The fixes created new bugs almost as often as they fixed old ones. The PTF tapes arrived with instructions to ''apply them at your own risk.''

Software's dirty little secret is that it has always been full of bugs. The difference is that in the old days, computers were kept in secured glass rooms and could be approached only by those initiated into the information-systems priesthood. We never had any illusions about software quality, but to the general public, computers could do no wrong. Now, with computers absolutely everyplace, the secret is out.

Richard Martin
Fairfield, Conn.


Computer software is doing a marvelous job throughout our economy. In almost every field, it reduces errors and promotes efficiency. It helps air-traffic controllers squeeze more aircraft into a given airspace, software-controlled machines build parts with close tolerances only dreamed of 20 years ago, and banks handle hundreds of millions of transactions with near total accuracy. Those errors that do occur are generally attributed to the users of software, not its programmers.

Let's put things in perspective. Of the 21 software failures cited in your sidebar, not one caused a loss of life. The outage of eBay for 22 hours was a non-event compared with the real problems we face in society. It's time to give software its due.

John V. Kjellman
Henniker, N.H.

What's Right About the Super Hornet

''The (not so) Super Hornet,'' (Industries, Dec. 13) depicts a limited view of a successful acquisition program. The Super Hornet's design process started almost 10 years ago, focusing on five key requirements: increased range, payload, bring back (the ability to land on the carrier with unused weapons or extra fuel), significant avionics growth as compared to the F/A 18C/D, and most important, affordability.

The F/A 18E/F Super Hornet successfully completed an extensive, three-year flight test program, during which the program remained on schedule and under budget while still meeting every key performance requirement. The Super Hornet's performance provides robust solution to the complexities of carrier-based naval aviation. I have the utmost confidence in seeing young Americans, including my son or daughter, fly this aircraft into combat.

Jeffrey A. Cook
Rear Admiral, U.S. Navy
Patuxent River, Md.

Why Philip Morris' PR Will Win Few Converts

''Philip Morris: Inside America's most reviled company'' (Cover Story, Nov. 29) noted that the impact of the Kraft boycott led by the corporate watchdog organization INFACT has been felt at the highest levels of the company. As internal documents show, Philip Morris executives and public relations advisers feared the boycott and urged cultivation of a reasonable, moderate image to defuse consumer action. Hence the suddenly conciliatory rhetoric of Philip Morris executives quoted.

There is no real attempt at dialogue, only a wave of new products from Philip Morris' spin factory, including its nine-figure ad campaign and its 11th-hour public conversion this fall to the idea that smoking is addictive.

The corporation's efforts to convince consumers of its wholesomeness focus increasingly on its Kraft Foods division. This was a losing strategy for RJR, which tainted its Nabisco subsidiary instead of Nabisco sanitizing its tobacco operation. An INFACT boycott of Nabisco helped persuade RJR to cut it loose.

While Philip Morris' new ''good citizen'' public relations strategy is likely to win few converts among angry consumers, it also has another target audience: the Supreme Court, now hearing oral arguments whether the Food & Drug Administration has the authority to regulate tobacco. With a new poll showing that 69% of Americans support FDA regulation, Philip Morris spin deserves no more traction with Supreme Court justices than it gets with the public.

Sangita Nayak
Organizing Director
INFACT
Boston

The Correct Stance on Ergonomics

''OSHA's ergo-rules: Business, hold your fire'' (News: Analysis & Commentary, Dec. 6) is on the money. Our small business has an annual workers' compensation safety inspection. We have made it a practice to implement all their recommendations. With our workers' compensation insurance premium at 0.5% of payroll, it is money well spent.

We have to move away from this us-vs.-them attitude when we review regulations. The Occupational Safety & Health Administration is wrong to think that all businesses have no concern for their employees. We know the cost of absenteeism. And the business groups would be just as wise to see what their members are doing before taking a defensive position.

We need to understand that we are all in this together. And we need to make our decisions, regulations, etc., based on what is best for all parties and not just a ''me first'' attitude.

John E. Falk
Executive Vice-President
Menlo Tool Co.
Warren, Mich.


Your message was that businesses should understand the benefits of an ergonomically designed workplace. This makes sense. Fewer injuries on the job save companies money and lost production time. Healthy employees are important ingredients in a well-run business of any kind.

However, you imply that, because of this, businesses should be supportive of OSHA's ''one-size-fits all'' ergonomics proposal. The American Apparel Manufacturers Assn. has spent the past 10 years on designing and implementing ergonomics programs tailored to the needs of the apparel industry. A sweeping federal program that contains ambiguous language is a hindrance, rather than a help, to our members as they strive to address ergonomics issues.

Larry K. Martin, President
American Apparel
Manufacturers Assn.
Arlington, Va.

''Vying for the fast track'' (Economic Trends, Dec. 20, 1999)

''Vying for the fast track,'' (Economic Trends, Dec. 20) should have noted that Connecticut and Oregon were among the 10 highest-rated states qualifying for ''honor roll'' status on the 1999 Development Report Card for the States issued by the Corporation for Enterprise Development.

''IBM: The biggest dot.com of them all'' (Cover Story, Dec. 13, 1999)

''IBM: The biggest dot.com of them all,'' (Cover Story, Dec. 13) incorrectly attributed $5.8 billion spent on general procurement to Carrier Corp. The parent, United Technologies Corp., spends that amount.

''Fuji: Beyond film'' (Special Report, Nov. 22, 1999)

The service offered by America Online Inc. in a partnership with Eastman Kodak Co. was misnamed in ''Fuji: Beyond film'' (Special Report, Nov. 22). It is You've Got Pictures.