Icon Solutions, NBC 4

SARAH
So Reese is crazy.

SILBERMAN
In technical terminology, he's a loon.

People. Do not follow in the path of Dr. Silberman, who has inspired derision in audiences worldwide.

A few years ago when I was consulting in New York I wrote:

It is hard to account for the silence of commentators, pundits, consultants, the press, in the face of the important information technology industry's stunningly poor performance.... Our complacent belief in the business acumen operating in our corporate institutions is contradicted by daily reality, by obvious mistakes, by deplorable results and wasted human resources.

[http://urielw.com/it.htm]

This was a big story that everyone knew. "My observations as a software consultant in many corporations suggest that easily a majority of software projects are colossal failures," I wrote in that essay. Yet in the alternate universe portrayed by the media, apart from a few high-visibility flameouts, the IT sector was performing brilliantly. Was everyone worried about their goddamn pensions?

Free your minds, people. Do not be slaves to the false reality concocted by your news media and punditocracy.

It's sad, isn't it, how action movies can be more authentic than the news media?

A loon, I? I think anyone with normal corporate experience in the last couple of decades who finds my observations implausible has rather a delusion problem himself. The emperor's new clothes are quite, quite non-existent. Virtually everyone working in the corporate world can tell a story of fiasco similar to the following, which a list member emails in response to my recent missives:

The old software worked seamlessly, it was more intuitive, AND it was more convenient for the essential task of typing. You could use tab to go to different fields, or you could use the arrow keys to work your way around the screen. There were also plenty of keyboard shortcuts. As a former programmer myself (amateur, in high school) I was appalled at the lack of attention to detail the new program was given by whoever coded it.

I kept asking whoever would listen why it was they were "upgrading" when it was so obvious that the new software was inferior to the old one. Since I always talked to the low-level people, they always agreed with me and shared horror stories of the time they had to spend learning the new software, being trained in it, adjusting old work habits, doing something for three hours when it used to take five minutes ... you get the idea.

Enron the norm? I can believe it.

And would it amaze you to hear that the NBC people (the main manager for Genesis, as well as two vice presidents) had no wish whatever to hear me tell them how the inaptly named "Genesis" was actually a sinking ship? (The hit movie Titanic had recently bobbed up in America's consciousness.)

While at NBC I met Hemda Fridman, a member of a team from another consulting company called TACT ("The 'A' Consulting Team") which had been hired to review Icon's work. She had authored a report to NBC on Icon's "Base Class Library," or BCL. Libraries such as BCL can be thought of as the software foundation upon which all project-specific software is built. While such libraries are intended to make programming quicker and easier, BCL had the perverse effect, as her report pointed out in some detail, of multiplying the amount of programming required.

"The declarative programming technique that is now a part of the class library unnecessarily forces the developer to code hundreds of lines of declarations," her report stated. "This not only dramatically increases the potential for mistakes, but also creates havoc when trying to debug."

The report had been submitted to NBC some time before I first met Hemda. Had some reorientation of the project therefore taken place? By no means. Icon was still blithely building all the software for the project on top of BCL, with NBC's blessing. And Hemda was rather leery of discussing the matter at all. I believe she was fairly alive to the value of her goddamn pension.

Since I had the chore of working within the BCL structure, I had already raised similar questions about it within Icon. That was not the way to my boss Rick Mohr's heart.

Further crippling our productivity was the technical development environment we worked in. In one of the status reports we were required to submit to NBC I wrote:

Three factors operate in parallel to hamper productivity:

- Using (slow) network copies of Windows and PowerBuilder instead of (fast) local copies.

- Using Windows 3.1 instead of Win95 or NT.

- Using 16-bit PowerBuilder instead of the less crash-prone 32-bit version.

With 15 team consultants working onsite for years and a regular drumbeat of pronouncements telling us how urgent everything was, NBC managers might have divined that hampered productivity might push costs in the wrong direction and adversely affect the project.

Did this lead to a transformation of the project's development environment? Not a bit of it. But I was directed to stop submitting status reports to NBC.

I suspect the phenomenon of television plays a part. (Come to think of it, the project failure at NBC might have been fitting.) I don't mean to exaggerate TV's role as bogeyman, but I think a yearning by corporate workers for the sanctuary from thought that TV offers must have had something to do with the rise of Microsoft Windows. Here was a way to savor TV's vacuity in the workplace. Microsoft must have understood this. How many millions of manyears have been spent modifying Windows wallpaper and mouse icon preferences? How many zillions of manyears go down the drain because people (even programmers!) use the mouse instead of the keyboard?

Maybe it's during Windows' rise to dominance that, with Microsoft leading the way, hype became everything, substance nothing. "GUI" (pronounced gooey) was such a pleasant term to bandy about, so delightfully intimidating in its effect on the losers who didn't know it meant "graphical user interface" (which in turn simply meant "like TV"), that anyone betraying partiality to old-time virtues like speed, efficiency, correctness, was seen as obviously a dolt or a cad.

Once software system interfaces became like TV, they became the center of attention, and no one had to think about logic. Now clueless managers would eagerly get involved in system design. The essential software development lesson we were taught in Computer Science 101 -- plan before you build -- was jettisoned as too harsh for managers. Gurus divining what clients wanted to hear came up with the kinder and gentler "iterative development" and "prototyping" methodology, which became standard practice. So there was no plan, and the stuff managers would look at during the various development stages was just a "prototype" that wasn't expected to work anyway.

But this is only conjecture. Maybe those who trace the disappearance of public integrity back to Kennedy's assassination are right. Someone else will have to figure out the history.

There is, admittedly, something uniquely impudent about the incompetence seen in the computer industry. Just a few days ago, the software company Network Associates turned up in the news. They are being sued by New York State for violating the First Amendment rights of consumers and journalists. Their software disks carry the warning: "The customer will not publish reviews of this product without prior consent from Network Associates Inc."

As Eliot L. Spitzer, the attorney general of New York, says: "Imagine if Ford had these clauses. Nobody would have been able to criticize the Pinto." (The Network Associates lawyer-lackey says "the company has not done anything wrong" and they only wanted to encourage customers to get the latest product version before publishing reviews.)

But as I noted in that essay years ago, the problem is broader than the computer industry, involving social trends carrying us "towards ever more superficiality, mindlessness and duplicity." That's why I think NY Times columnist Paul Krugman's end-of-era prophesy is unduly optimistic. It'll take more than Enron and temporarily tightened accounting standards to roll this back.