(FORTUNE Magazine) – IT HAS BEEN JUST OVER six years since Carleton S. Fiorina, now 50, burst upon the national stage--and we will acknowledge straight out that FORTUNE played a role in putting her name in lights. Back at the takeoff point, in 1998, she was the accomplished, high-energy president of Lucent Technologies' core business. FORTUNE, publishing its first-ever ranking of the 50 most powerful women executives in the U.S., put her smack at the top of the list. A line heading the accompanying article said, "It may surprise you that our No. 1 woman is someone you've never heard of."

That was the end of Fiorina's quasi- obscurity. Less than a year later she was named CEO of Silicon Valley's famous Hewlett-Packard. That post, making her the head of a company that now has $80 billion in revenues and is the 11th largest in the U.S., vaulted her to a level in the corporate hierarchy that a woman had never before attained. For years she had a near lock on first place on our annual list of women executives (though she was edged out last year by eBay CEO Meg Whitman). Her reputation bloomed, heading toward rock-star celebrity. She became one of the few businesspeople identifiable by her first name: She was just "Carly." Totally poised, she gave countless speeches; she became the only woman extant who never had a bad-hair day; she was the subject of endless rumors that she might move on to politics.

But celebrity, as everybody knows, isn't an achievement in itself. Beneath the public image are the yardsticks against which executives are--and should be--measured. So it is right to ask whether this whirlwind has succeeded. And inevitably that question must be answered in two parts. First, under the only lens that matters, did the famed merger that Fiorina engineered between HP and Compaq produce value for HP's shareholders? Second, with that merger nearly three years past, is HP in shape to thrive in its brutally competitive world?

The answers are no and doubtful.

LET'S BEGIN WITH that deeply controversial HP/Compaq merger, carried out in May 2002 after a proxy fight of red-state/blue-state ferocity. Winner Fiorina moved briskly into integrating the two companies, brushing aside warnings she would fail. Time passed, and the company did not implode. True, its earnings progress was inconsistent and sometimes the stuff of headlines. And the stock was worrisome, since it went nowhere. Still, a general, vaguely based opinion began to develop, often reflected in the press, that this might be one merger that you couldn't condemn.

To anyone thinking that way, don't. This was a big bet that didn't pay off, that didn't even come close to attaining what Fiorina and HP's board said was in store. At bottom, they made a huge error in asserting that the merger of two losing computer operations, HP's and Compaq's, would produce a financially fit computer business. The irrefutable evidence on how wrong they were is contained in the two companies' own merger proxy, which precisely laid out the healthy operating margins that the combined company expected to be earning in its 2003 fiscal year (HP's ends Oct. 31) on its computer operations. The margins weren't earned then, and in 2004 they weren't either--not by a long shot. Only the prodigious, money-coining strength of HP's star business, Imaging and Printing (better known as "printers"), has kept the company looking respectable.

The fundamental and overpowering problem here is that HP's shareholders paid $24 billion in stock to buy Compaq and in exchange got relatively little value. In fact, so little value was secured that accounting rules could force HP to write off a chunk of the $14.5 billion in goodwill assets it set up on its books after the deal. That would be a noncash charge--comparable, in a junior-grade way, to charges taken at AOL Time Warner after its debacle of a deal--but a write-off of goodwill at HP would say as clearly as anything can that, financially, this merger has been a lemon.

Moreover, in the way that mergers work--and this is the true, if seldom recognized, crime of the deal--HP's issuance of roughly 1.1 billion shares to Compaq's shareholders, to be added to HP's existing 1.9 billion, means that HP sold about 37% of its assets to the Compaq crowd. Among those assets is that gem of a printer business, whose 40% 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% of the printer business. After the merger, they owned only 63%.

So--before we hear Carly Fiorina's side of this argument--stand up, Walter Hewlett, and take a bow. It was Hewlett, son of HP founder Bill Hewlett, who as a director fought the merger and lost the proxy fight. It was crazy, he said, to in effect sell part of the printer business just so HP could move more deeply into the achingly competitive, heavily commoditized computer world. Initially, Hewlett did not declare what he thought HP should do instead--just sit with the business portfolio it had, or what? Ultimately, he suggested that HP might spin off the printer business to its shareholders. Meanwhile, his opponents derided him as a cello-playing dilettante pushing off-key plans.

After he lost the proxy fight by a whisker, Hewlett contested the vote in court. An angry HP board retaliated by kicking him out as a director. When Hewlett then lost his lawsuit, he soon began to sell his HP stock. In time, most of the 110 million shares that he, his father's estate, a Hewlett trust, and the Hewlett Foundation owned were unloaded.

That would make this cellist a good judge of stocks as well as mergers. Ever since September 2001, when the Compaq deal was announced, HP shares have been a dead-in-the-water asset. Recently, at about $20 per share, HP was 13% below its price just before the merger news hit the market. That doesn't compare well with anybody that counts, not even archrival IBM, down 6% for the same period. On the flip side, HP's chief competitor in printers, Lexmark, rose by an impressive 60%; and HP's nemesis, Dell, absolutely thrived, leaping almost 90%. The Lexmark record is especially telling given that it hints at how well HP's printer business might have done had it been spun off into a stand-alone company.

HP's stock market blahs, combined with the inadequacies of its computer profits, makes the merger hard to defend. Indeed, one HP director, Robert Knowling Jr., CEO of the New York City Leadership Academy, displaying a frankness rare for board members when talking to the press, makes no effort to do that. The merger, he says, "hasn't met the board's expectations or the management's." But Knowling is nonetheless strong in his support of how Fiorina has been running the company. That in itself is revealing about the merger: If he thinks she's done well with what she's got to run, and it's still far from producing the profits expected, then there must have been something wrong with the strategy that created what she's got to run.

You will not get such an admission out of Carly Fiorina, to whom FORTUNE talked at great length while reporting this article. True, she concedes that neither the stock nor profits are where they should be. But that's not because HP's acquisition of Compaq was wrongheaded, she says. It is instead, she claims, the result of a more extended downturn in IT spending than many expected and of weaknesses in "execution" that she is driving to fix.

Meanwhile, she thinks back on the post-bubble days of 2001, when the Compaq acquisition was booting up, and argues that drastic action at HP was essential. The critical problem was buckets of red ink spilling from computers. And significantly, says Fiorina, "neither the market nor the organization understood the difficulty HP was truly in." She wishes now she had found some way to be more "transparent" then about HP's problems. That would have prepared people for the merger. But candor, she maintains, would have also further damaged the company, hurting employees' confidence, sales efforts, and probably the stock price. So she kept quiet. This backward look, of course, can be viewed as somewhat suspect (or even "self-serving," as one institutional investor puts it), in the sense that it argues Fiorina then had to do something dramatic, like a huge merger.

In what appears to be rationalizing, Fiorina and her top executives often look right past the profit disappointments and find great satisfaction in the company. Fiorina sees HP today--this "FORTUNE 11" company, as she sometimes calls it--as a well-provisioned, intensely global, end-to-end corporation, marching along with not only its printers but also $55 billion in revenues from the whole IT panoply, including PCs, servers, storage, software, and services. With this gear, she says, the "enterprise" part of her business, which sells to large corporations, has become a viable competitor of IBM's--which once it was not.

Certain of Fiorina's managers also definitely like what they're running more than they used to. The combination of HP's and Compaq's PC operations produced a business that, while only marginally profitable, is still stronger than the losers the two companies had separately. Most of all, combining the two computer services businesses created a more solid operation than existed at either place before. That's a big reason Ann Livermore, the HP executive vice president who runs the Technology Solutions Group, which includes services, says, "I don't know anyone who has a thing to say against the merger."

Well, hello again--here we are. The problem with all these structural virtues, good as they may look on a sales call, is that they cost HP shareholders $24 billion to get, resulted in the bargain sale of 37% of the printer business, and aren't producing decent profits. In other words, the merger may have improved HP's status, but it did so only at an indefensible cost and without producing a company of merit.

Stay with us as we de-layer just how much HP makes. Bottom line, according to Generally Accepted Accounting Principles (GAAP), 2004 profits were $3.5 billion. That's a dull 4.4% of $80 billion in revenues. Nor does it stack up well against assets of $76 billion (the ratio is 4.6%) or stockholders' equity of $37.6 billion (9.3%). When the next FORTUNE 500 comes out with its performance rankings for such key measures as return on equity (in which the 500 median is likely to run about 14%), this huge company will place way down the list.

HP's 2004 profit figure, however, was reduced by certain charges (including restructuring items related to the merger) that the company prefers not to think of as operating expenses. So in reporting to Wall Street, HP adds back these charges and arrives at a non-GAAP profit figure of $4.1 billion, which is roughly equivalent to operating earnings after taxes. The pretax operating earnings are $5.3 billion. And of that--get ready for a heavyweight and a lightweight in the ring--the printer operation accounts for more than $3.8 billion, while the rest of the company tags along with about $1.5 billion.

Even then, some outsiders hypothesize that HP is internally allocating every cost it possibly can to printers so that it can keep profits down to a level that will not suck new competitors into the business. Dell made its entry anyway and is currently getting an insider's look at margins. It has told at least one analyst that it finds printer consumables (inkjet cartridges and the like) so remarkably profitable that it believes HP has to be making more in its printer division than it is acknowledging.

In any case, it is the non-GAAP profits that Wall Street analysts run with, dwelling almost always on the twists and turns of today. In 2004 these included a thudding earnings miss for HP in its third fiscal quarter and a rousing bounceback in the fourth. In the minds of analysts hangs also the everlasting possibility that the "execution" Fiorina is promising might actually produce a long string of consistent quarters in which printers continue to shine and all the computer businesses suddenly wake up and do well.

But no one spends much time remembering that even HP's best quarters in computers are so far away from what the company forecast in its merger proxy (and in marketing "road shows" as well) that it is embarrassing. This proxy was sent out by management and the board in February 2002 (a point at which no one could claim ignorance of 9/11's effects). The document was officially filed with the SEC; it contained some "cautionary" boilerplate language about risks but was nevertheless precise in its expectations; it was the last plea to shareholders about to vote. And it said that integrating the two companies' operations and facilities would rationalize the product line, produce billions in savings, and create computer businesses with respectable operating earnings.

Ironically, the integration part of this plot worked fine: Products were successfully meshed, and billions in costs were cut. But there is a saying, "If a thing is not worth doing, it is not worth doing well." When the integration was artfully completed, the hardware parts of the computer operations--the source of nearly $40 billion in revenues last year--stood there naked, barely making profits. And the $13.8 billion computer services business, though certainly a far better earner than hardware, not only didn't make its 2003 goal but backslid in 2004.

Here is what the proxy statement forecast for operating profit margins in 2003 and how far away from these targets HP was even in 2004:

In dollar terms, of course, these percentage shortfalls translate into missed billions. Had HP made the margins on its computer businesses in 2004 that were forecast for 2003 in the proxy, and had all other things been equal, pretax operating earnings last year would have been almost $2.4 billion higher than they were. That would have increased the pot from $5.3 billion to about $7.7 billion. The incremental dollars would have been taxed, but HP pays relatively modest taxes because it manages to earn a large portion of its money in low-rate jurisdictions overseas. So imagine, for the sake of argument, that $1.9 billion of the increment had come down to the bottom line. That would have raised HP's return on equity from the 9.3% it recorded to 14.4%, which would probably leave it just about matching the 500's median.

When you have flubbed a forecast so completely, you may also have blanked out memories of the deed. Fiorina herself needed to be reminded where these forecasts that FORTUNE was questioning her executives about had come from. Talking to us, Duane Zitzner, who ran the Personal Systems Group--that's PCs--for three years before retiring in mid-January, spoke of the 3% forecast for his business as an out-there goal simply to be fought for. "Rome wasn't built in a day," he says, "I am extremely proud, and I've told the board and financial analysts this, of what we've accomplished. Absolutely, beyond doubt, to my grave, I believe we have done a great job in this business in lifting it to where we have." Recall next that the lifting has taken margins to 0.9% against the 3% forecast printed in the proxy.

All this leaves open the possibility that the forecasts were just so much proxy-fight froth, never meant to be believed. But that would be legal daredeviling, and one HP director, Verizon president Lawrence Babbio, is quick to say the forecasts were legitimate: "Merger proxies should be looked at seriously." His own viewpoint is that a "harsh" operating environment has prevented HP from meeting its targets.

Then again, one HP executive wrote FORTUNE an e-mail presenting a weirder explanation. "While investors are certainly a very important constituency," the e-mail said, "the truth is they are not the reason we did the merger. The reason we did the merger was to be a stronger, more capable partner to our customers."

Harder to rationalize is the fact that HP may have to acknowledge the merger's shortcomings on its books--by writing off a portion of the goodwill assets it has on its balance sheet. Oversimplified, here's the accounting problem: HP's payment of $24 billion for Compaq left it having to delineate, on its balance sheet, how much of that amount was assignable to specific items, including tangible assets and one intangible asset of note, the value of the Compaq name. That exercise took care of $9.5 billion of the purchase price; the remaining $14.5 billion became goodwill. And under the rules, it was assigned to specific HP computer businesses, called "reporting units."

The difficulty now for HP is that two of these reporting units, PCs and enterprise storage and servers, are earning only meager profits. Therefore, the "economic value" of the goodwill assigned to them is suspect. Under accounting rules, this suspicion must at least annually be resolved by comparing the "fair value" of these businesses with their carrying value, including goodwill, on the books. If the fair value is less than the carrying value, the goodwill is impaired and must be written down. And how do you measure the fair value? You estimate such bedrock items as future cash flows and theorize what you could sell the businesses for to a willing buyer.

In this exercise of very large judgment calls, some Wall Street analysts have an implicit opinion: They argue today that there is no value in the price of HP's stock for computers. That is, they think the stock would be selling where it is were HP simply a printer company. That implies that HP couldn't sell its computer reporting units for anything, which would absolutely say there's bloat in goodwill. Maybe the analysts are overstating things. (IBM, after all, did just unload the bulk of its losing PC operation on Lenovo for $1.75 billion.) But here's an opinion from one merger and acquisitions authority, Scott Baxter, who teaches the subject at Columbia: "When I look at how little HP is making on the computer assets it bought from Compaq, I have a very difficult time seeing that much, if any, economic goodwill exists. If that's the case, though this would be a hard pill for HP to swallow, it should be writing off goodwill."

Clearly HP and its auditors, Ernst & Young, have rejected that proposition so far. But Robert Wayman, HP's respected chief financial officer, says that the issue is patently there for the company to contend with, especially in the struggling enterprise business: "There's no question that if profitability doesn't improve, there will be an impairment charge." He makes the case that it wouldn't matter much, given that it would be just a bookkeeping entry. But responding to the thought that such a charge would both rudely expose the failings of the Compaq merger and resurrect talk about the bitter proxy fight, he says, "Sure. It would be in the press."

Even today, echoes of the proxy fight resound in the belief of some Wall Street analysts, particularly Steven Milunovich of Merrill Lynch, that sooner or later the company will be broken up into its parts. Milunovich argues that the split-up companies would sell for more than the $20 a share that HP is commanding currently. One plank in his platform is that the market would react well to each company's having to stand on its own feet, meaning, for instance, that printers would not be forced to subsidize PCs.

There appear to be some institutional investors that believe in at least some variation of this argument. One of particular interest is Private Capital Management (PCM), of Naples, Fla., a unit of Legg Mason and a "value" investor so successful that it has for the past ten years bested the record of that firm's famous Bill Miller. Last year PCM made itself a large HP shareholder by acquiring an estimated 38 million shares, recently worth about $760 million. PCM doesn't publicly discuss its investments. But its president, Bruce Sherman, did say some weeks ago that there was "sum of the parts" thinking at work.

The difficulty for any investor hoping for a split-up is that Carly Fiorina, and apparently the HP board, don't want one. True, Fiorina caused a stir at an HP analysts' meeting in December by saying with some asperity, in answer to a question about breaking up the company, that the board had considered the idea three different times--and rejected it. The press leaped on the comment, probably too eagerly, because it is likely that at least two of the three times she had in mind preceded the merger. Only if the directors had looked hard at a breakup recently would there be reason to think it is a serious possibility.

And Fiorina is not commenting on that: "I am not going to describe when the board deliberates on what," she said to FORTUNE a couple of days after the analysts' meeting. She then added, "What I was trying to convey is that the board has been deliberate, thoughtful, complete, and unanimous over and over and over in believing that keeping this portfolio together is the best option for shareholders."

That's what she herself has claimed to believe again and again. Her pride, which she is not short of, is bound up with somehow making this company work, even if not by the definition detailed in the merger proxy. Some people fault Fiorina for having rapidly moved out of jobs at two previous employers, AT&T and Lucent, without having really "finished" her work. Those critics think she might speed from HP also, perhaps by riding the rumors into some kind of Republican post. But her HP job seems unlike the others in that her managerial reputation is thoroughly and publicly on the line--in a company that hasn't succeeded--and would be badly tarred if she walked out. If the HP board stays with her, she is not a candidate to exit.

That is, of course, the other matter of constant interest: Just how solidly is the board behind Fiorina? A few weeks ago, one secondhand but relatively credible report reached FORTUNE that the board is split over Fiorina. The two outside directors that we managed to talk to, though, Babbio and Knowling, firmly denied that was so. We did not get calls back from director Richard Hackborn, a strong-minded, retired HP executive who built the printer business. Hackborn played a key role in hiring Fiorina in 1999, and she is thought to need his continued support. Without it, her job might be imperiled. But don't forget that this board made this merger and has its own pride on the line.

WHICH BRINGS US to part two of our inquiry: assessing the future of Hewlett-Packard. As director Knowling says, "We're off the diving board." Okay, given that the merger is history, can this company compete from here on in?

One problem inherent in the company's challenge is what HP isn't. The history of tech shows that companies that dominate their pieces of the business are the big moneymakers--companies like Microsoft, Intel, and in its heyday, IBM. Leaving aside printers, HP doesn't dominate anything. Beyond that, it is imprisoned in its structure. It must deal with both the relentless competition in computers and its own particular need to battle on two fronts, against both IBM and Dell--a "precarious" position, says Knowling. Fundamentally, computers for HP are a Big Lousy Business, the very hardest kind to handle. This BLB also recalls Warren Buffett's famous adage: "When a management with a reputation for brilliance tackles a business with a reputation for bad economics, it is usually the reputation of the business that remains intact."

So the CEO's job at HP is exceptionally difficult, perhaps beyond the ability of anyone to handle well. Still, the latest proxy available, for 2003, shows that Fiorina was paid around $7 million in cash and options. It's therefore fair to hold her to a high standard. FORTUNE asked a dedicated supporter of Walter Hewlett in the proxy fight whether some other CEO could be doing better than Fiorina in handling HP's problems. Calling that "a good question," he proceeded to go all around the barn. "She's a very, very smart, competent, talented executive," he said. But then he hammered on some faults, mainly her inability, he thought, to admit she was wrong in getting herself trapped in computerland: "She cannot bite the bullet and say, 'We lost.' Other businessmen"--slip of the tongue--"can do that and move on. She can't."

In the end, unless the board suddenly swerves into a fast U-turn, Fiorina is probably going to get more time to struggle with HP. That will undoubtedly leave her performing in her usual style, applying enormous energy to the job. FORTUNE got a look at her calendar. George W. Bush would pale at her nonstop schedule and need a short rest to recover from just having glanced at it. She is forever on the road, traveling all over the globe to see customers and very often heads of state. At this marketing work, she is first-class.

Fiorina is also extraordinarily and amazingly articulate in all settings. Most other CEOs grope at least occasionally for words; she never seems to. Her talent for talk naturally solidifies her reputation for brains, though it would seem that one of her staff members recently stretched things a bit by saying, "I wonder what it feels like to walk into a room and know always that you are the smartest person there." But there is also a downside to her verbal skills in that what she is saying so fluently sometimes wears thin with people, coming off as short of substance, overly aspirational, and lacking in spontaneity.

Well beyond the merger proxy, Fiorina has not served herself by habitually engaging in financial forecasts that meet analysts' misbegotten calls for "guidance." As a new CEO she got caught flat-footed twice with her forecasts when the bursting of the tech bubble destroyed earnings. Much more recently, in December 2003, she told analysts that HP thought it could increase earnings per share by more than 20% annually. That is an extravagant ambition for any giant company (see "The 15% Delusion" on fortune.com), let alone one trucking along in a tech industry often described as "maturing."

The folly of big promises became terribly apparent last summer (just a short time after Fiorina gave a totally upbeat speech at Allen & Co.'s big Sun Valley conference). Early on Aug. 12, Fiorina shocked analysts--some of whom had been dragged from their West Coast beds by 4 A.M. phone calls--by announcing that earnings per share for the July quarter just ended would be nearly 23% below the figure analysts were expecting. HP's stock immediately dropped by more than two points to about $16.

Three problems, all in the enterprise business, underlay this profit jolt. First, European results suffered because HP mismanaged some of its dealings with the resellers and distributors that the company works with. Second, in the U.S., project managers failed to get a new order-processing system online as scheduled, which reduced shipments of enterprise computers. Third, HP's storage business didn't meet its profit goals. Storage, intrinsically a business of lovable margins, has in fact given HP major headaches for some time.

The largest significance of the third quarter is that Fiorina summarily fired--within hours--the head of enterprise sales, Peter Blackmore, and two of his senior executives. The speed and visibility of the move caused the tech industry to view the firings as an unnecessarily cruel public hanging, and criticism of Fiorina flared. She argues back, saying that a hanging it wasn't and that in truth these people were treated, in the usual HP manner, with "respect, candor, and compassion." But she also says that some of the European problems had been ripening for a time and that these three very senior executives clearly "dropped the ball." She believes in accountability, she says, so the men are gone. Reached by phone, Blackmore declined comment because, he said, he and the company had mutually entered into a nondisparagement agreement (to which he appears to be sticking a little more strictly than HP).

And thus it was that three more executives left HP, which under Fiorina has lost --via firings, resignations, and retirements--a noticeable number of its top people. Here's one indicator of what some people call the "brain drain": Of the 11 "direct reports" that Fiorina had in October 2003, one died and five others are gone. Michael Capellas, CEO of Compaq when the merger was done, left early. Former HP employees can also, without much effort, list another 20 well-placed executives who have departed, a few of them going to competitors Dell and storage company EMC.

Fiorina does not agree, naturally, that there's been a brain drain. In fact, she believes that one lesson she's learned while running HP is that she should have moved more quickly in ejecting certain people. Smartened up now, she says, "I would have done them all faster. Every person that I've asked to leave, whether it's been clear publicly or not, I would have done faster."

As 2005 starts, HP has no earnings forecasts on the table. One reason for staying clear of them--as if there weren't always good reasons--is that the company, along with corporate America generally, is about to run into the world's most famous accounting requirement: the expensing of stock options, which is slated to begin midyear. HP has always issued lots of options. They were initially the means by which Bill Hewlett and Dave Packard ushered other employees into ownership of the company. So they became part of the celebrated "HP Way"--the caring, "Bill and Dave" culture that sometimes leads people to call HP "the church of Silicon Valley." The culture has created anomalies at HP: dividends, for one, and a defined-benefit pension plan in a tech world that hardly knows the term. The fabric of HP has been tested in recent years, particularly by the terminations of armies of people. But Fiorina says the culture is vital to the success of the company and must be preserved. So she expects to keep spreading options deep into the ranks.

The company concedes, though, that the new options-expensing requirements will ax earnings by an amount that will be "material." As for what that might mean, HP's recently filed 10-K says that if the rules had been in place in 2004, profits would have fallen by over 20%, which by anybody's definition is material.

Whatever the hit turns out to be, it is a downer that HP can't avoid. The mighty task for the company will be overcoming that vexing problem--and a whole lot more. Can it find rich, new veins of profit to mine and at the same time hold its competitive ground with the businesses it has?

On one front, HP is pushing its digital technology into consumer electronics (though that's a competitive field if there ever was one). HP has, for example, by far the bestselling media center PC, which is meant to replace the living room TV and stereo. But the market for that product is currently small, and in certain other consumer electronics areas, HP's efforts have actually earned it some scorn. Trying to tap into the music craze, it last year began selling Apple iPods under the name "iPod by HP." The notion struck some people as an affront to HP's honored reputation for innovation. In any event, the iPod by HP isn't slated to be one of those coveted veins of profit. According to Apple, HP accounted for just 7%, or 320,000, of the 4,580,000 iPods shipped during the Christmas season.

Searching for rich lodes in the enterprise space, and willing to accept some startup losses, HP has been aggressively buying small software companies. That's a move recognizing one of the verities of selling to IT departments: After software comes hardware, and after hardware comes services. The services referred to in this chain of thought are primarily the customer support kind, capable of being lucrative.

The other branch of this business, less of a payer, is managed services, in which companies outsource their IT departments or business processes. IBM is the champ of this highly fragmented industry, but the Compaq merger has helped HP become a worthy contender. TPI, a consulting organization that tracks outsourcers, says that last year HP got about 7% of the contracts above $50 million in size, which put it No. 2 on the list. But IBM was a runaway No. 1, at 15% share.

The trouble with managed-services contracts is that they are black boxes whose profitability can be ascertained only down the pike. Witness, for example, EDS's infamous losing contract with the Navy, which the company heralded at the deal's inception in 2000 but has since wished it could just deep-six. New contracts, in addition, sometimes levy heavy startup costs of their own. That's one reason HP's services fell in profits last year.

An extreme difficulty about assessing the progress of a computer company in managed services or anywhere in the enterprise world is that so much of the evidence is anecdotal. P&G, for example, loves the work that HP has done for it on a ten-year, $3 billion outsourcing contract, a true prize for HP when it obtained this business in 2003. But the chief information officer of an even bigger company than P&G says that he cannot see that HP has the global might--as IBM does--to take on his outsourcing work. Nor does he find HP competitive today in storage products. And this CIO said he did not know the senior HP salesperson who was covering his account--an observation that obviously disturbed Fiorina and Blackmore's replacement, Michael Winkler, when they had it repeated to them.

In the end it all comes back to the bottom line, where HP's enterprise business, as we know, has been floundering. Sure, the buzzwords are out at HP: It's striving to deliver a better "total customer experience" and to increase its "share of the wallet." But those are things every other computer company is trying to do as well. Meanwhile, all the sellers operate in a world in which cost-rabid CIOs are increasingly viewing their servers and PCs as commodities and playing off their suppliers against each other. In two minutes of talk recently, one bigtime CIO expressed his admiration both for Fiorina--"She's dynamic! A lightning rod!"--and for a favorite cost-cutting strategy of his that is guaranteed to drive executives like Fiorina nuts. The strategy leans on "reverse auctions," in which buyers solicit bids on their website for large quantities of machines, driving their price down sharply.

In almost everything that HP does today, there's the shadow of Dell. It is a macho competitor in the server business; it endlessly rolls out PCs; and now it is nastily there in the printer business. Making this newest assault, Dell had in mind defense: It has publicly said that it wished to prevent HP from freely using its printer profits to make life difficult for Dell in PCs. But just because there was preemption at work here doesn't mean that Dell isn't totally determined to do well in printers. It has clearly suited up for battle.

The HP executive most in the way of this onslaught is Vyomesh Joshi (called "V.J."), head of the printing business and the newly named boss of PCs as well. Making a really standout managerial move, Fiorina promoted V.J. to run printers in 2001 from another spot in the division. The business was the uncontested market leader at the time, but he has since increased its operating profits by a stunning 84%. Describing what she saw in V.J., Fiorina says he was "incredibly passionate" about his work, a quality that comes through when he talks.

Some of his work today involves pushing into copiers, the ground of Xerox and Canon. And in the enterprise business, he is eager to build sales in the broad world of corporate printing. A statistic he loves to quote is that digital printing, which is the turf of HP, has only 4% of the total printing business in the U.S, which remains an analog world. He aims to raise digital's share.

HP's traditional printer business has a razor-and-razorblade look: The plot is to get the printer into home or office, and from there on make hay selling consumables. In this business, V.J. is banking on his division's acknowledged record as an innovator. But competition comes from many angles. On one front he's facing vendors who would like to hijack his consumables business by selling cheaply priced remanufactured or private-label ink and toner cartridges, or simply by refilling HP cartridges with ink. V.J. charges this bunch of sellers with purveying unreliable goods.

When the subject gets to Dell, V.J. leans on emphatic opinions. First, he stresses how completely HP dominates the printer business: It shipped 47 million printers in 2004, compared with Dell's expected five million. That ratio will change, of course. But V.J. thinks tensions will develop between Dell and Lexmark, its main supplier of printer products, as the latter begins to resent how much of the consumables business Dell is siphoning off (and in fact, analysts say there are tensions). Right or wrong, V.J. also contends that Dell's system of selling by mail will not work well for printer cartridges.

In the end this may turn out to be a battle between product innovation, which is what HP expects to bring to the party, and business-model innovation, which has made Dell, the fast shipper of low-priced products, what it is. Let's just say one thing: No matter the outcome, the aggressive presence of Dell in the market cannot be good for HP. Business doesn't work that way.

IN THE MIDST of all the competitive pressures bearing down on her, and in the struggle of managing the unwieldy company she created, Carly Fiorina sometimes talks as if she sees a vision all her own. She hauls it out in the opening lines of internal speeches, articulating her goal of making HP "the world's leading technology company." The ambition is a curiosity rouser because it implies that she has firmly in mind what company right now holds that title. But that turns out not to be true. There isn't any one company that fits the bill, she says. It could be IBM one place and some other company elsewhere.

But then, you wonder, if she doesn't know who the leader is now, how would she know whom HP has to pass and when to claim victory? Oh, well, classify this as one of those aspirational themes that Fiorina wafts skyward now and then. The utility of the idea seems small, anyway: Whatever it is, HP is not right now close to being the world's leading technology company. ■