SAN MATEO (03/27/2000) - As time goes by, things that were once considered difficult always become easier. Back in the 1980s, setting up a local area network took luck and experience, whereas in the 1990s the intricacies of deploying a suite of ERP (enterprise resource planning) applications nearly consumed us. Today, both tasks can be accomplished using procedures that have withstood the test of time, making these once-daunting tasks seem downright manageable.
As we enter the 2000s, the biggest challenge for IT is the constantly evolving nature of Web sites. Just when you've got a handle on what makes your site work well, up pops a new set of required functions that throws your existing assumptions out the window.
But like all technologies before it, Web site construction and management will get easier. Future Web sites may consist of little more than a GUI connected to a set of Web services. This notion, called e-provisioning by some, is an outgrowth of the application services model. But instead of outsourcing an application, Web sites will provision different services to deliver an integrated suite of functions to their customers.
For example, instead of building a customer service application, a Web site will link to a service from a company such as Agillion. Instead of buying an e-commerce application, a Web site will link to a service from somebody such as Escalate; and instead of buying a tool for personalization, a Web site will link to a service provided by Personify.
Ultimately, you should be able to build a Web site without building a single back-end application. Of course, that's a few years away because none of these Web services are readily aware of each other and even fewer have well-defined paths for sharing data. But the promise is there, and the final pieces are starting to drop into place. For instance, once XML schema definitions become rich enough to serve as an interface to a variety of applications, sharing data across applications will become a standard feature of every program written.
Once that becomes established, the only real hurdle will be to find a way to automatically discover what services are available on the Internet. That's when directories will finally live up to their full potential. As a technology in search of a killer application, directories have been perceived as something you should deploy, rather than something you must have to run your business.
As the investment community says, there are only two kinds of products: vitamins and aspirins. Vitamins are good things you're supposed to take but never do. Aspirin is something you can't live without because you have a crippling headache.
Directories are about to make the conversion into an aspirin-type product, because efficiently finding services on the Web gives everybody a headache.
Consider, for instance, all the phantom digital exchanges that are currently on the drawing board. Right now, the only way to tightly integrate two partners on a digital exchange is by putting a small army of developers on a plane to hardwire a set of applications into the exchange. Given the number of developers in the world, we should finally finish getting everybody tied into all the exchanges in the world just in time for the Y3K crisis.
So what has to happen is that as people build applications, they need to register them in a directory that can then be searched across the Web by other applications. And once found, integrating that application with another program will only require sending an asynchronous message to call a component interface. And, of course, if you easily find a service on the Web, there's little reason to build or buy yourself, which in turn means it should become a whole lot easier to build and maintain Web sites.
Michael Vizard is editor in chief at InfoWorld.