Microsoft Bats a Thousand at PDC

If Microsofts professional developers conference were a baseball game, Microsoft would have gone three for three: good base hits, though none of them home runs. The portfolio of technologies shown at the conference in Los Angeles late last month was substantially standards-based, appropriately compatible with what has come before and capable of changing the way that enterprise developers work. Those were all solid responses to the challenges that I pitched to Microsoft before the event.

The PDCs extended opening keynote emphasized openness, with one demo after another using ordinary ASCII editors—not proprietary tools—to write declarative XML or straightforward .Net code. The resulting applications combined rich-client power with transparent access to Internet resources.

The choice of Amazon.com, a Linux shop, to perform a major keynote demo was strategic. Amazons developers made it clear, in a subsequent conversation with eWEEKs Darryl Taft and me, that they enjoyed the opportunity to deliver their standards-based Web services in a more customer-friendly way to "Longhorn" clients. Im sure that Microsoft will be pitching its .Net managed code on the server, as well as Longhorn technologies on the client. I was pleased to see at the PDC, however, that enterprise adopters arent faced with an all-or-nothing proposition.

Before the PDC, I urged attendees to look for solid grounding in nonproprietary standards for all Microsofts future initiatives (see Labs-Eye View, Oct. 27, eWEEK). I encouraged them to seek specifics, placing on Microsoft the burden of proof that its forthcoming technologies are worth two years of developer preparation to put them to work—and that theyre also sufficiently defined that it makes sense to start that preparation today.

Microsoft met this challenge, overall, with demos that bore out the comment of publisher Tim OReilly during our conversation later that week. He observed that what we used to call a "personal computer operating system" is now essentially just a device driver for one element of the more interesting application platform. The "Indigo" unified messaging model, a critical but mostly invisible part of the Longhorn portfolio, represents a large step in that direction. Developers writing Indigo code are using the PC, the Internet, and the many devices and processing nodes available via the Net in the same way that Visual Basic 1.0 developers gained unified control of the Windows client environment.

In addition to my three-for-three assessment of Microsofts overall PDC performance, Im inclined to award bonus points to keynote speakers Bill Gates and Jim Allchin. Both achieved a tone of candor, including frank admission of past errors and technical disappointments, that probably makes it easier for PDC attendees to entrust these guys with the future of their companies and their careers. Gates poked fun at his own delayed recognition of the Internets importance; Allchin stood in front of professional symbol grinders and fired up a Unix hackers vi editor window to sling a few lines of without-a-safety-net code. This is the kind of direct connection with no-nonsense coders that retains and strengthens their loyalty.

The goal of a PDC is to gain the undivided attention of the people whose choice of a platform is a self-fulfilling prophecy of its success—but Microsofts crucial relationship with developers has never been challenged as it is today and as it will be over the next two years. Open-source economics will continue to be compelling; alternative platforms, such as Suns and Apples, will continue to entice enterprise buyers with their emphasis on reducing costly complexity and providing a secure foundation for e-business.

The Longhorn canvas offers developers a paint-by-numbers scene of capability and convenience in delivering applications. Recent history suggests that developers want that combination, more than anything else that other platforms or other doctrines might offer. Those who pay the enterprise IT bill, though, must look at more than initial cost of development. It will be their task to balance developers enthusiasm for Longhorns efficiencies against the lower cost of deploying an open-source alternative.

Technology Editor Peter Coffees e-mail address is peter_coffee@ziffdavis.com.