The Observer: Why Do We Take This Abuse?

When did we become such weenies? And why have we allowed that to happen?

Those were the two questions I walked away asking myself after a thorough reading of InformationWeek's superb package on software quality in its May 20 issue. (That a pore-baring photo of Bill Gates appeared on the cover of said issue is one of life's delicious ironies and one we'll deal with later.)

Most striking to me were two findings from InformationWeek's crack research team: first, that 89% of the 800 IT managers contacted say they've experienced software-quality problems within the past year that resulted in higher costs, lost revenue, or both; second was that less than 70% of the IT managers seem to hold software vendors responsible for or able to repair those problems.

The same InformationWeek research shows that rather than asking vendors to fix a problem, a vast preponderance of the IT managers surveyed are opting to have their own programmers fix problems, would rather trash or simply stop using a flawed product, or--in a mind-boggling finding--will suffer with the poor-performing product simply because it's already been paid for.

Think about that for a second. In what other conceivable market (other than, perhaps, barbering) would aggrieved consumers not expect as a matter of course that a provider of a product or service make good on a bum deal? Yet that is precisely the case when it comes to software.

Could a restaurateur, a carmaker, a hotelier, or a newsstand operator in effect bilk customers and have those customers simply walk away without balking? Better still, would any of their customers simply agree to cook their own meals, fix their own cars, make their own beds, or go without a newspaper even after the meal, the car, the bed, and the paper had already been paid for?

When and how did the culture of cowering consumers take hold when it comes to software and, seemingly, only software? A seed for that development comes, I think, from a mind-set that's typified by a quote that appeared in Mary Hayes' Quality First story. In her piece, she quotes a spokesman for a software vendor as saying his company's product is "the most complicated software made, period. And installation is very much akin to rewiring a nervous system."

I won't cite the spokesman's employer in the interest of fairness; his company is hardly the only vendor to engage in that sort of pomposity and self-aggrandizement born of magnifying the scope of the products it provides and the dizzying array of skills required to accomplish same.

But likening linking machines together via the language of zeroes and ones to rewiring nervous systems stretches an analogy to the point of absurdity. Heart surgery is complex; building craft that weigh many tons, lift hundreds of people miles into the air, and land safely is a complex endeavor. Granted, writing software code is part art, part science, and the results are breathtaking when it's done well, but compared to any number of products and services we encounter daily, its complexity pales. Most importantly, practitioners of the coding arts aren't deities.

Longer ago than I care to admit, I spent an extended period of my career covering what were then known as systems houses--outfits that wrote specialty vertical-market software that they'd bundle together with a particular vendor's hardware and resell to users. In the late '70s, these were hot shops and the software these often smallish systems houses wrote was regarded by customers as one might a particularly dangerous solvent--phenomenally useful, but to be handled with extreme care, lest it blow up on you.

Those customers were right to react that way. This was specialty coding, done by a handful of people at small firms. Their software didn't come from companies with tens of thousands of employees, global presences, and, most tellingly, elaborate ways and means of assuring quality. Yet we users apparently continue to look upon software in precisely the same way that those early-adopter companies did 25 years ago: If the software doesn't work, it must be because I've done something wrong.

Finally, the irony of Bill Gates being InformationWeek's cover boy on the week when software quality was the subject at hand. I leave you only with this thought: InformationWeek found that when it asked 734 IT managers which classes of commercial applications produced the greatest numbers of bugs or errors, operating systems and desktop-productivity applications (such as Word and Excel) were separately cited by just under 70% of the respondents as being the worst culprits. No other category came within 10% of these dubious totals. I rest my case.