服务组合是构建面向服务、松耦合、高适应性的应用系统的主要途径.现行服务组合技术大多只适于构造需求明确、业务流程可预先定义的情况.然而,现实中存在大量需要边执行边探索、“摸着石头过河”式的问题求解形式.针对这类需求,结合网络化科研协作中的实际问题,提出一种最终用户可探索式服务组合的方法,重点讨论了探索式服务组合编程模型、服务推荐等关键问题.此外,从使用者角度对所提出的组合方法及其实现机制给予了定性分析和评价.
Service composition is a prevalent means to construct service-oriented, loosely-intcgrated and agile applications. Current service composition technologies are mostly developed for the situations where business requirements are definite and business processes can be predefined. However, there exist many problem-solving cases in which people are not totally clear in advance how to compose applications, but have to act in a try-and-error manner. A novel approach to dealing with the challenge is proposed. Focuses are on the programming model supporting exploratory service composition as well as a dynamic service recommendation mechanism. The proposed approach is evaluated from the users' perspective.