Sure, this merger has flaws wide enough to drive a truck through, and it is by no means the industry-defining deal HP's chief executive claims. But the merger just might let a once-great company stagger back to its feet again. The corporation that helped set the pace and define the values for Silicon Valley probably won't reprise that role anytime soon, but it just might become a pretty good company in a few years — if Wall Street and the families of its founders, some of whom are contesting the deal, get out of the way and let management run a business.

HP is suffering from a number of problems no deal will cure. The company will never fill the void left by the passing of engineer founders Bill Hewlett and Dave Packard, whose offices remain preserved in the former Page Mill Road headquarters in Palo Alto, Calif., like shrines to the respect and deep loyalty of those who worked under them.

No one can recreate the wide-open space for innovation and profits that has been forever sucked out of the computer industry with the rise of the Intel and Microsoft monopolies. And HP managers are frank about the many missteps of the past decade, which have left them in less than a strong competitive position today.

That said, there's a future for a merged HP and Compaq, and plenty of smart and dedicated employees are waiting for the insightful orders that lead to it. The key is investing in what real core innovation a computer company can still command. That may require letting go of something else.

CEO Fiorina should give up her aspirations for a broad, branded-product portfolio of business and consumer products that range across commodity areas such as consumer PCs and digital cameras. The money saved from slugging it out in some of those markets could be reinvested in servers and information appliances. Indeed, if HP is to succeed for the long haul, it must invest more in engineering tomorrow's products than in maintaining its brand on today's commodities.

HP/Compaq revenues would lead computer OEMs: Profits, R&D would trail

The executives themselves stake a big share of the merger's future to their server businesses. One of the secrets of success here is HP's IA-64 deal with Intel. In compensation for defining that architecture and helping bring it to market, HP gets a significantly lower price on the IA-64 processors than any other system maker. If you spread that lower price across all HP and Compaq servers, it starts to get interesting enough to make Dell sweat.

And if the merged company is able to maintain that deal into the far future, when the IA-64 trickles down to desktops, the story gets even better. This, no doubt, is one keystone in the thinking behind the merger.

But the executives have their work cut out for them. Almost every server in their combined product lines faces an architectural transition to IA-64. HP uses its PA-RISC chip, the Tandem Non-Stop systems use MIPS, Compaq's high-end line uses Alpha. There are plenty of painful bridges to build and cross, and each one likely will cost customers.

The merged company can turn this to its advantage by designing best-in-class interoperability tools to smoothly unite its own disparate systems and to go after Sun's and IBM's customers as well, leading the industry to a future polyglot computing utility.

Servers are one of the rare spots in computing that still promise profits. The merged company needs to capitalize on that by leveraging its design capabilities to the hilt. The decision earlier this year to ax a group of 100 ASIC designers who helped build the chips powering HP's high-end Superdome system smells of a quick way to cut costs — at the expense of an important limb of systems design expertise. The merger cannot afford many such missteps.

The integration team needs to nurture the VLSI design group in Fort Collins, Colo., that by all accounts has been doing an excellent job crafting the McKinley IA-64 processor. Lay on whatever fuel Compaq has to add to that fire. The more value the merged company can bring to highly optimized and reliable servers, the better it can make hay with those low IA-64 prices.

And a bold public statement about plans to continue designing microprocessors and ASICs for the long term would help win the hearts and minds of the smart engineers the company so desperately needs.

Think Linux

Systems software is another core opportunity the merger must not miss. The company should become the best Linux developer on the planet, designing all the interfaces, plumbing and tools that OS so urgently needs. This is the space where Microsoft won't tread. Don't leave it to IBM or Red Hat.

That goes double for what are euphemistically called "access" products. Stop building Windows CE/Intel StrongARM handhelds. That is a business with little profit and less future.

Instead, try to acquire Handspring, which is taking the PalmOS into the strategic communicator space. At the very least, take a license to the PalmOS or BeOS, and begin designing on a platform with more potential for innovation than what Redmond allows. In short, get out from behind your bunker and lead the way to the pervasive-computing future.

While HP works multiple sides of the software fence, it should do the same in microprocessors. HP has a great partnership with Intel in high-end CPUs. How about striking one with Advanced Micro Devices on the low end? Maybe HP could help AMD deliver best-in-class, low-power or low-cost processors that could power its notebooks, low-end servers and desktops.

Finally, the hardest task — rearchitect an engineering culture suited for the 21st century. That's not a job for a former AT&T sales and marketing executive (Fiorina) and a former IT and supply chain manager (Compaq's Capellas). Find a strong engineering champion. Be bold, and recruit Pat Gelsinger from Intel, John Kelly from IBM or Anant Agarwal from Sun. Then give that champion a charter to reignite innovation in a broader, deeper way than merely pasting the word "invent" on a logo.

Bill and Dave pioneered an urbane culture of consensus and work/life balance at Hewlett-Packard that set the tone for Silicon Valley. But now they are gone and the Internet era has arrived. Today's startups — and HP's coopetition partners Microsoft and Intel — have taught the Valley a new lesson about rapid innovation, dogged focus on execution and the need for continual change.

HP needs to brand that fire on its engineers — after it convinces them they are the most important people in the company. I'm not sure the current management of HP or Compaq really believes that. But they had better start. Otherwise, this merger will become little more than a temporary stay of execution, as the great big company battles Dell down to a fraction of a percent of profit margin, with Intel and Microsoft laughing all the way to the bank.

Related stories:

HP looks to create another MPU peak

Promising efforts follow HP's technology slide