An analysis of frequent ways of making undiscoverable Web Service descriptions

Abstract

The ever increasing number of publicly available Web Services makes standardcompliant service registries one of the essential tools to service-oriented application developers. Previous works have shown that the descriptiveness of published service descriptions is important from the point of view of the algorithms that support service discovery using this kind of registries as well as human developers, who have the final word on which discovered service is more appropriate. This paper presents a catalog of frequent bad practices in the creation of Web Service descriptions that attempt against their chances of being discovered, along with novel practical solutions to them. Additionally, the paper presents empirical evaluations that corroborated the benefits of the proposed solutions. These anti-patterns will help service publishers avoid common discoverability problems and improve existing service descriptions.