Down the road, look for services to change the way the entire computing infrastructure functions p:bigbite/multi-title –> BEYOND PROVIDING A lower-cost infrastructure for application integration, the advent of Web services lays the foundation for a slew of changes in enterprise computing that will manifest themselves during the next several years. The first profound change will come with a shift in the way people think about building and deploying applications. In fact, in the future, people will not build applications as we know them today at all. Instead, they will build digital services. The subtle difference between an application and a service is that most applications built today are assumed to be stand-alone entities. But in the age of Web services a new style of enterprise computing will emerge because instead of building applications, developers will be asked to build services where the ability to integrate one service with another using protocols such as XML, SOAP (Simple Object Access Protocol), and UDDI (Universal Description, Discovery, and Integration) will be assumed and required. Ultimately, developers will create services that leverage other services to create fabrics and even tapestries of apps that were unforeseen by the people who built the initial service. None of this is going to happen tomorrow. But during the next few years this software-as-a-service concept will become the dominant model for development. This is a topic we at InfoWorld will explore in-depth this week at our Web services conference in San Francisco, which counts Eric Rudder from Microsoft, Adam Bosworth from BEA, James Gosling from Sun, and David Winer from UserLand Software among the keynote speakers (see www.infoworld.com/nextgen ). Once the applications and operating systems are able to dynamically discover one another, discrete pieces of a distributed computing environment can be broken up into its logical components. Furthermore, these components can be linked together to create a grid that will allow users to instantly call on any subset of the overall environment. In the near term, that would mean that instead of having to wait to install additional servers to increase capacity, an IT organization could just bring additional servers on-line that would dynamically learn about the computing environment in which they are to play a part. Longer term, you can see how different subsets of the distributed grid can be optimized for specific parts of a service. Once this type of infrastructure is in place, it starts to become apparent that the industry as a whole will shift to a utility model where organizations are billed for usage of computing resources, rather than each organization trying to build and maintain its own datacenters. We think the people primarily charged with turning these visions into reality will be the CTOs who are the backbone of any serious IT operation. Our upcoming CTO Forum conference scheduled for April in San Francisco will include featured speakers such as John Seely Brown from Xerox and Tim Howes from Loudcloud to help explore these developing trends (see https://ctoforum.infoworld.com ). Hopefully we’ll see many of you at these and other upcoming InfoWorld events in 2002. Software Development