HomeIndustry Commentary

Cottonwood now or hardwood too late?

Like Tweet Pin it Share Share Email

160 years ago we were a nation of farms and small merchants, in which both people and information traveled no faster than when Julius Caesar overthrew the Roman republic.

A mere 45 years later, the trip from the east coast to California, formerly a half-year ordeal, was a routine two week journey; information traveled coast-to-coast in seconds; and giant corporations were considered persons by the law.

And you thought the Internet transformed the world? Compared to the transcontinental railroad, the agent of these transformational changes, the Internet hasn’t come close.

Take time to read Stephen Ambrose’s Nothing Like It in the World, which chronicles the process through which Abraham Lincoln preserved the union from an east/west division just as surely, if less dramatically, as his leadership of the Civil War preserved it from secession by the southern states.

Whenever you read history you learn something useful about the present. Among the many instances here are foreshadowings of both Enron and a current IT controversy.

The Enron foreshadowing? Both the Union Pacific Railroad and Southern Pacific Railroad companies created separate construction companies. The construction companies, owned by the railroad company board members, siphoned off huge amounts of capital from the railroad companies, enriching their owners while leaving both the employees and other railroad shareholders holding the bag. Sound vaguely familiar?

But that doesn’t directly affect you as a working leader of information technology. This does, as a metaphor for problems you deal with every day:

For the Southern Pacific Railroad, hardwood railroad ties were in short supply. To avoid construction delays, it used cottonwood ties in much of its initial construction instead, even though they would last only a few years. Bad engineering? The company desperately needed cash flow. Had it waited until it could acquire enough of the “right” railroad ties, it would have run out of money before finishing construction.

This philosophy explains a lot of what we’ve experienced with Microsoft operating systems over the years. Time-to-market has usually trumped internal engineering, to the despair of many in our profession. Was this decision wise? The question is, for whom? Certainly, the earlier Microsoft is able to ship product the earlier it makes a profit. What — you thought Microsoft was an altruistic enterprise, putting your best interests ahead of its own? That isn’t how business works. From its perspective, generating revenues earlier was worth the impact on its image with respect to quality.

Sometimes bad engineering is the best engineering — a valuable lesson. An example even closer to home:

Quite a few readers griped about my endorsement of Fire Controlman Derrick Thomas’s homegrown database, which reduced the time needed by the U.S.S. Benfold for each Persian Gulf ship inspection by two and a half hours. Thomas should have worked with an IT professional to build it right, I was told … after all, it was probably an undocumented mess, and who supported the system once he left the ship?

Had Thomas had taken this advice, he’d have waited until the Benfold returned to port, then submitted his request. Had it ever emerged from the priority queue the Benfold would have received the system a year or two later.

It’s cottonwood now or hardwood too late: A fleet of destroyers’ worth of capital and wages, times 2.5 hours per ship inspection, times a few hundred inspections, all spent because end-users shouldn’t build systems by themselves?

The quality of Thomas’s code is unknown. But the quality of his business judgment?

Outstanding.