The Carly Fiorina era at Hewlett Packard ended Feb. 8. Much of the coverage has focused on her charismatic style and her rapid rise from senior sales executive at AT&T to the highest-profile woman executive in corporate America.

Fiorina's removal was unexpected, though not shocking. Shares of HP stock dropped 50 percent during her six-year reign. (The Nasdaq dropped 20 percent in the same period.) Still, many analysts expected that she would be given at least till the end of 2005 to execute her vision of a merged HP-Compaq as a serious competitor to Dell in PCs, and to IBM in corporate systems integration and outsourcing.

HP had a profitable 2004, primarily because of the profitability of its printer and imaging business. But the computer manufacturing division, though now nearly as large as Dell in revenue, remains far less profitable. The new giant cannot compete with Dell's low-cost model any more than HP and Compaq did separately.

At Dell, which began as a mail-order operation in Michael Dell's University of Texas dorm room, a PC is built only after a customer has ordered it. Components are purchased from suppliers on a "just-in-time" basis, cutting out the depreciation of parts and PCs that plague the rest of the industry.

The final straw in Fiorina's departure might have been Fortune magazine's Feb. 7 cover story, "Why Carly's Big Bet Is Failing."

The article, by veteran business writer Carol Loomis, used a simple but devastating investment analysis -- commonly associated with Warren Buffett, with whom Loomis has close ties -- to illustrate how much shareholder value HP lost by acquiring Compaq in 2002.

"The fundamental and overpowering problem here," Loomis writes, "is that HP's shareholders paid $24 billion in stock to buy Compaq. ... [In effect] HP sold about 37 percent of its assets to the Compaq crowd. Among those assets is that gem of a printer business, whose 40 percent market share [according to IDC] makes it one of the great franchises in the world. To sum up the damaging mathematics: In the beginning, the old HP shareholders owned 100 percent of the printer business. After the merger, they owned only 63 percent."

What could HP have done to be profitable? From the unromantic mind-set of 2005, HP might have been best off pursuing a niche strategy, investing in profitable businesses such as printing and imaging, and retaining the profitable instrument division, rather than spinning it off as an IPO into Agilent. This would have involved gradually stepping away from the low-margin PC market, as IBM has done.

More than ever, the long-term trends of the IT industry are becoming clear. During the past 15 years, computer hardware has imploded as an independent, profitable industry, one segment at a time. The mainframe manufacturers went first, followed by mini-computers, supercomputers, PCs and workstations. Each was devastated when its products became commodities, plastic shells that contained advanced chips and software from other companies.

The flat margins of nearly all hardware vendors, aside from Dell, suggest that HP's problems in the PC market are not those of execution, but of continuing to farm once-fertile but now-barren markets.

Under Fiorina, HP sought to replicate IBM's success in transitioning from a struggling technology vendor to a profitable services provider with deep technology expertise. And HP has had several big wins in corporate outsourcing -- such as the $3 billion, 10-year contract signed in 2003 to manage Procter & Gamble's IT and back-office systems.

But the rapidly expanding outsourcing industry is also experiencing growth pains, with profit margins being driven down by offshoring and the structure of long-term contracts. HP has potential as an outsourcer, but a more difficult path to success than IBM.

IBM's success is arguably attributable to its positioning itself as an IT "conglomerate," the General Electric of its industry. Much as GE uses its Financial Services as the "glue" that stabilizes and adds value to its capital equipment businesses, IBM uses its long-term outsourcing and services contracts to insulate profitable product lines, such as enterprise software and mainframe hardware, from short-term market pressures.

HP does not compare with IBM in the depth of its R&D and product portfolio, or have IBM's decades-long record of managing complex, enterprise systems for global Fortune 500 companies. Nobody does.

What will HP do now? There is still value in the printing business and potential in services. But the future of the HP's computing businesses is bleak.

In the 1990s, HP, Compaq, DEC and Sun were wildly successful computer vendors, and PeopleSoft and Informix were blue-chip software firms. Aside from HP and Sun, they are no longer independent firms. Neither are the Minnesota computer giants -- Sperry, Control Data and Cray.

As much as anything, HP's quandary is a comment on the unforgiving nature of the information technology industry. Retailers that fail usually fail over the course of decades, after a long series of missteps. IT vendors, in contrast, can fall once and never recover, declining at as drastic a pace as they once grew.