One explanation for what’s wrong with two key facets of IT—the state of the CIO’s job and the enterprise software market—came together when I read a recent help-wanted ad from the Internal Revenue Service. Two rather odd specifications for an associate CIO position—responsible for modernizing the IRS’ IT systems—caught my eye: The candidate cannot be a current or former IRS employee (or a federal employee of any kind), and his or her tenure in the position is limited to four years.

Think about that for a minute. No one with an inside understanding of how the IRS functions can apply, and the winner of this dubious sweepstakes must be out of there in what amounts to a fraction of the time it would take to get anything reasonable accomplished, such as fixing the IRS’ IT mess.

Meanwhile, the enterprise software market is marching to a vastly different timetable. If you look at where the top applications and infrastructure vendors are heading, much of what constitutes the technological future of the industry—SAP’s NetWeaver, Oracle’s Fusion, Microsoft’s Project Green—won’t be fully baked for a good three years. All the cool stuff we’re all waiting for—service architectures, process management, even supposedly down-to-earth concepts like radio-frequency identification—will need a similar time frame before they get real. Which means that any successful candidate for the IRS post will be heading for the door just when things start to get interesting.

But it’s even harder to know what the IRS candidate would be able to do with all of these new technology toys. Pretty much everything significant that’s rolling off the enterprise software assembly lines these days is deeply business-centric, and requires, if nothing else, a profound understanding of how a business’ processes should be handled by enterprise software. The outsider who takes this job, whose insight into the IRS need only extend to accurately filling out a 1040 form, won’t be burdened by any such critical understanding.

This isn’t just an issue for the IRS: It’s endemic to IT departments worldwide. The problem of long-range planning and short-range tenure has never been more acute. The need to define the applications architecture of the future requires an unprecedented far-sightedness: Innovation in the coming decade will require the definition—and execution—of a solid enterprise software architecture that businesses can use as the basis for future automation and efficiency. This marrying of IT and business needs strong insider business smarts in order to succeed, independent of technological expertise.

It’s ironic that the vendors are taking such a long view at a time when too many IT managers and CIOs are, relative to the evolutionary forces at work in the software industry, lame ducks. There has to be some way to give the CIO the longevity to plan for the long term, a longevity that requires vision and ensures follow-through. It’s easy to plan for a future you won’t be around to be part of. What’s hard is to plan and live to reap the benefits of success. As the IRS will undoubtedly find out, to do otherwise only guarantees failure.