Web services represent the third attempt to rebuild enterprise IT on the standards-based foundation conceived by Tim Berners-Lee. To realize that vision, vendors and IT pros must resist some familiar temptations and avoid some traditional mistakes.
The first Web was a tool of research, and it has succeeded beyond all expectations in putting facts at our fingertips. In the process, though, it has also provided split-second access to harmless nonsense and malicious deceptions, dimming the luster of that first attempt to redefine our way of life and work.
The first lesson learned from the Web should be that quality matters at least as much as convenience. The Web can serve a valuable role as a global source of both the profound and the inane, but an enterprise-oriented Web needs tools such as the source authentication of IPv6 at the infrastructure layer and the professional analysis of a competent research staff—whether in-house or outsourced, but in either case more discriminating than Googles PageRank algorithm—at the content layer to refine what it delivers.
The second Web was the tortured adolescence of the dot-com boom, filled with the arrogant zeal of the teenager trying to prove the old rules no longer mattered. Predictably, brilliant visions were greatly outnumbered by unrealistic and ill-founded attempts to build castles in the air. The moral of that second Web is that sound business models create the opportunities for using new tools, not the other way around.
This brings us to the emerging third Web, which will be a storehouse of business logic as well as data. In the third Web, the proven disciplines and business models of the enterprise are finding new mechanisms for doing the things that have cost so much, and taken so long, to learn to do well.
We see the transition of Web services, from intriguing demonstration to genuine tool, in offerings such as the Mainframe Express Enterprise Edition product launched this month by Maryland-based Micro Focus International Ltd. The product gives developers new power to combine the availability and integrity of mainframes with the management flexibility and developer productivity of Web services. At months end, we hope to see more such offerings at the JavaOne Conference in San Francisco.
But it would be a sad repetition of past errors if IT vendors seized upon the success of Web services as their latest excuse for locking customers into product stacks—perversely building proprietary structures from standards-based components, for example, by devising proprietary XML schemas for their applications and documents.
IT buyers need resolution to avoid being seduced by the accompanying convenience of one-stop shopping: Fulfilling the potential of Web services requires a long view of future needs for testing tools, security infrastructures and other building blocks for which only open architectures can enable best-of-breed integration.
eWEEK is interested in your Opinion. Send your comments to eWEEK@ziffdavis.com.