How high are our expectations for the software and data infrastructure on which our computers, our networks, our very businesses depend? Not high enough, according to Peter G. Neumann, chief scientist at Stanford Research Institute's International Computer Science Laboratory ("House of Cards"). We take much too much for granted: guarantees of network and Internet security, promises of reliability on the part of software developers, assertions of functional completeness on the part of implementers and integrators. The result: We put up with systems that don't work the way we'd hoped they would, increasing security problems that slow down the growth of e-business and a porous Internet that's still waiting for the big blackout ("Internet Chernobyl?").

How high are our expectations for the successful completion of IT projects? Judging from the results of this month's survey on project management ("Project Management"), they're not very high at all. Even though just 10 percent of our survey respondents' projects come in both on time and within budget, the respondents awarded themselves an above-average rating on their ability to manage projects. In talking with CIOs about why projects go wrong, we heard lots of discussion about the difficulty of working with project owners to set and maintain priorities. But IT executives need to take responsibility, too, for managing the expectations of other executives about the realities of completing projects on schedule and within budget.

If you're faced with a truly problematic project, perhaps the best solution is to kill it, and kill it early. In this month's Whiteboard ("How to Kill a Troubled Project"), Gopal K. Kapur draws up a strategy for deciding when to stop troubled projects. Among the criteria: Does it fit with business strategy? Does it have the right sponsors, and have the intended end-users bought in? The answers to any of these questions may lead to the decision to end the project, but the criteria can also shed light on how to bring such projects back to life.

One CIO who has done a first-rate job of meeting the very high expectations his colleagues hold for him is this month's Catalyst, Don Haile, CIO at Fidelity Investments, the country's largest mutual-fund company ("Balancing Act"). Haile manages the day-to-day workings of one of the world's most complex IT organizations—Fidelity will spend around $2.2 billion on IT in 2001, about 20 percent of its annual revenue—while initiating such hugely challenging projects as Fidelity's implementation of XML in hopes of streamlining its middleware development program. At the same time, he has pushed his company hard on the development of new technologies, including wireless and voice-recognition systems. Every company should think like Fidelity, which, despite its categorization as a financial services firm, steadfastly thinks of itself as a technology company.