6.7 Conclusion

The way Web services are envisioned , companies acquire publishing information from a registry (such as ebXML or UDDI), build an implementation that meets the requirements, and then publish their services to the registry. Other companies then discover a service listed in the registry, comply with the established standard, and conduct business. Once agreed-upon standards and rules of engagement have been met, applications can take over, removing human interaction from the overhead of transacting business. In this model, using existing infrastructure, applications may exchange information regarding services that a company wants to provide and consume . Moreover, applications may continue to poll the registry for other applications (companies) that provide identical services under more favorable terms.

XML is about enabling remote access to services through a standardized approach to laying out data and service requests using a lingua franca for machine-to-machine intercommunication. With such an alphabet, we can get on with the task of asking what kinds of interactions people want their machines to undertake for them. We can determine how machines might be constrained to behave so that they will tend to achieve their masters' objectives within appropriate resource, time, and security constraints. And finally, we can determine appropriate ways of dealing with outages, violated expectations, delays, inconsistencies, etc., that would normally make machine-to-machine interactions too brittle for everyday users to rely upon for highly personalized (low volume, specialized, customized) services.

The next chapter provides information on agents and how they can be of use for the me-centric environment.