A watershed event for the testing industry occurred in December 1999, and I'm not talking about Y2K. For the first time--that I am aware of, anyway--the sorry state of test practices was openly discussed in a national forum.
The cover story of a mainstream newspaper was actually focused on the commonplace failure of computer systems and the reasons why they occur. (See www.usatoday.com/life/cyber/tech/ctg838.htm.) This article was revealing for several reasons, and all of them are important to the present and future of software quality.
But the ultimate victory was that in most of the examples the real reason for the failure was admitted to be inadequate attention to testing. Hallelujah.
Now that we're well into the new century, it's easy to put Y2K fears behind us. Frankly, I was worried that after New Year's Eve 1999, all of the attention being paid to testing would dissipate and we would return to the old days, when software quality issues were practically invisible.
But this newspaper article, amazingly enough, announced "Y2K is dead" in December--before we knew that Y2K problems were really behind us. The author had the guts to make this statement because the relative cost of what was deemed "everyday failures" is a staggering billion a year, which is, coincidentally, the sum total that was spent on Y2K in this country for the entire preparation effort! In other words, there was a global hue and cry for years about the risks of a once-in-a-millennium event, when we are wasting as much money each and every year on problems that occur each and every day.
One hundred billion dollars a year. Think about it. By anyone's measure, that's a whole lot of money. It's incredible when you think about how many test departments are struggling to eke out the most meager of budgets and protect the stingiest of schedules, while all along the stakes are this enormous.
How could an elephant of this size be lumbering through our economy and just now be noticed? Well, it has a whole lot to do with denial.
For industry insiders, especially those who have made a career in testing, the news that systems are regularly slapped together and rushed into production prematurely, without adequate testing, is not news. It's standard operating procedure. For the rest of the world, however, this comes as a shock.
Yet if you follow the media exposure about software failures and the money being spent accordingly, the only thing that appears to be worth testing is the volume capability of your Internet site. Boring old functional errors in the tiresome applications, such as payroll, inventory, and shipping, that actually operate most corporate giants just don't seem to rate any real drama--or time or money.
Granted, in most cases, companies don't rush out a press release when they bungle their own operations, but sometimes they must. Hersey Food Corp., for instance, was forced to explain to stockholders why they had a 12% decrease in 1999 third-quarter sales and a 19% drop in net income from third-quarter 1998. It seems that a complex shipping system that was not thoroughly tested crippled deliveries during the critical Halloween season. But even a case like this, which cost millions in revenues and more in shareholder value, was not picked up as a hot story by either wire services or the networks.
It's show time
What will it take before companies stop wasting billions of dollars rushing systems into production without adequate testing? My theory is that this will only happen when these previously hidden systems are finally exposed to the outside world. Just as companies can't deny failures that escape to the Internet, neither will they be able to ignore those that affect other companies.
The fast-approaching business-to-business integration across companies that is sweeping the economy will likely provide the catalyst that reveals the shaky foundations of the most mundane but critical applications that do things like manufacture, inventory, and ship goods. Once the partners, shareholders, and other stakeholders understand that these errors are preventable, tolerance for deficient testing will evaporate and accountability will set in.
Not to be a ghoul about it or to wish public humiliation on anyone, but I actually look forward to the day when testing is no longer treated as an option or an orphan. A day when people don't glaze over and move away when you tell them what you do for a living, when inexcusable blunders are truly no longer excused. Until then, watch this space for news that might otherwise be buried. //
Linda Hayes is CEO of WorkSoft Inc. She was one of the founders of AutoTester. She can be reached at firstname.lastname@example.org.