OPINION
Bill Gates and the Asteroids of Doom

I've been thinking a lot lately about the future of computing -- and that, of course, has to include Microsoft's (Nasdaq: MSFT) future. Microsoft's own vision of its future seems clear. The company sees a transition to a secure Windows OS environment, continuing strength in Office and server applications, and development of a complementary set of business applications -- ERP/SCM and CRM -- as its route to the future.

That vision is compelling. If achieved, it would allow a business user to license everything from one source, use well-known applications like Word and PowerPoint as live interfaces to production data, and provide operational security in terms of high reliability and a relative invulnerability to malicious action.

There are customer costs to this that go beyond accepting the end of Wintel software development as a cottage industry and the establishment of a near monopoly in business applications to match what Microsoft now has in Office applications.

The Backward-Compatibility Equation

For example, the security solution will require the customer to accept hardware identification and a further layering of network protocols to cover encryption, recognition, authorization and so on. More interestingly, use of the Office tools as data interfaces to the business applications will make things easier for smaller users but much harder for large, multisite organizations that will no longer be able to use Unix database servers to glue together their applications.

Microsoft's bet is, reasonably enough, that customers will accept these costs. After all, improved security is a big carrot, and the small system demos will get slicker long before anyone notices that staffing requirements are doubling for bigger businesses that attempt to deploy the technology.

On the other hand, they're not betting that customers would forgive them for giving up backward compatibility. That's too bad, because the dream probably can't be delivered without that sacrifice. In other words, I believe Microsoft could deliver on its vision of Web-enabled, secure, single-source business computing if it started from scratch and built everything on Itanium, but not if it tries for backward compatibility.

By trying to build on past mistakes, the company is making the task exponentially more difficult, empowering virus writers and other attackers, and tying its own scalability to that of the x86 architecture. In theory, Microsoft could still pull it off, but my bet is that the company will just muddle along on this, abandoning the dream to introduce bits of Longhorn technology as they become available, making a big marketing play out of each small advance, and generally overselling relative to actual deliveries.

Asteroids and Doom

This, of course, is exactly what Microsoft has done all along. It's true today, for example, that Windows 2003/XP is a better product than Windows NT 4.0 was in 1998, but it's also true that essentially nothing of Microsoft's mid-1990s vision for reliable, self-correcting software made it into this century. That's probably what someone in 2008 is going to say with respect to today's promises too -- that the 2008 products are better but don't reflect plans from 2003.

In saying this, I'm assuming, obviously, that Microsoft will still exist as a software maker. That's not a given; in fact, it's only barely more likely than unlikely.

There are quite a lot of things that could drive major change, starting with a voluntary breakup into at least three operating companies and the return of substantial cash assets to the shareholders. Two, however, stand out as potential company killers -- the asteroids of doom.

Both are unlikely, but either could extinguish the company and quite possibly land its founders and key managers in jail.

Destroying the Entrepreneur

To understand the first and least likely of these possibilities, ask yourself what happens if you take the brightest people right out of university, offer them extremely well-paid and highly prestigious jobs in their chosen fields, and then stultify all their efforts in a huge bureaucracy while telling them their efforts are leading the field.

What happens is that you destroy their entrepreneurial potential. It took 35 human-years to build and start selling Opera, but more than 10,000 human-years to build Internet Explorer 3.0. Think about what that means as an indicator of the management structure and productivity in the Microsoft organization. Dick Pick built the first releases of his combined OS and application environment while working full-time for the Army; Microsoft promises to release something, someday, and commits a billion dollars to the first go-round.

MIcrosoft is famous for hiring the best and brightest right out of college. Remember Brooks and The Mythical Man-Month? When Microsoft puts 9,000 people on a project calling for a team of four or five, what do you think those people do all day?

Eight to ten years out of school, most of these Microsoft brain-trust recruits have a spouse, a mortgage, an assumed right to the kind of lifestyle US$125,000 per year buys, solid training in bureaucratic survival, a vested interest in the status quo, and a big-organization view of business that makes starting and running an entrepreneurial business unthinkable. That's why Seattle isn't a kind of Silicon Valley north. There are more high-potential university-educated computer science graduates there than in any other single location outside Palo Alto -- but fewer spin-off businesses than you find at minor research centers from Santa Clara to Rockcliffe.

Investors and Marketeers

Look around you. See many ex-Microserfs breaking new intellectual ground? Starting exciting entrepreneurial companies in high-tech fields? There are a few, yes, but that number is disproportionately small, and even those few are more often investors and marketeers than owner-managers.

Suppose it turns out that Microsoft has been destroying these kids on purpose as a way of keeping them from forming competing companies, joining the open-source crowd or working with Sun. That's an asteroid of doom because no one would want to buy from, or be associated with, a company whose owner-managers are so depraved that they preach a cult of youth to sell their products while cynically destroying generations of potential competitors.

The other possible asteroid is both less grim and far more likely. Suppose it turns out that Microsoft has a long-established pattern of duplicitous conduct with respect to real and potential business partners. Imagine the company offering, for example, business and financial help to potential competitors, using that to gain control of the technology and then either deflecting those rivals as competitive threats or helping them bankrupt themselves by creatively withholding promised cash, services or market support.

The Sendo Equation

That's very much what Sendo alleges happened to it -- and that case will be heard in a Texas court this summer. In different ways, this kind of conduct has been alleged with respect to RealNetworks' current lawsuit, the whole Netscape debacle and the change at Citrix from the NCD strategy of offering Windows access to Unix users to the Microsoft strategy of offering Unix access to Wintel users.

The current legal actions on this are relatively small and can be stalled in the courts -- merit notwithstanding -- for years if not decades. The era-ending asteroid threat here is from disgruntled HP (NYSE: HPQ) and DEC shareholders, not Sendo or RealNetworks (Nasdaq: RNWK). If it turns out that at least half of what Sendo alleges is at least half true, then simultaneous action through the courts and the SEC could bring the Microsoft era to a snappy end.

How likely is an asteroid impact of this kind? Overall, I doubt it's even 30 percent. But it's certainly high enough to warrant some serious consideration.

---

Paul Murphy, a LinuxInsider columnist, wrote and published The Unix Guide to Defenestration. Murphy is a 20-year veteran of the IT consulting industry, specializing in Unix and Unix-related management issues.

---