%0 Journal Article
%T The Semantic Automated Discovery and Integration (SADI) Web service Design-Pattern, API and Reference Implementation
%A Mark D Wilkinson
%A Benjamin Vandervalk
%A Luke McCarthy
%J Journal of Biomedical Semantics
%D 2011
%I BioMed Central
%R 10.1186/2041-1480-2-8
%X SADI - Semantic Automated Discovery and Integration - is a lightweight set of fully standards-compliant Semantic Web service design patterns that simplify the publication of services of the type commonly found in bioinformatics and other scientific domains. Using Semantic Web technologies at every level of the Web services "stack", SADI services consume and produce instances of OWL Classes following a small number of very straightforward best-practices. In addition, we provide codebases that support these best-practices, and plug-in tools to popular developer and client software that dramatically simplify deployment of services by providers, and the discovery and utilization of those services by their consumers.SADI Services are fully compliant with, and utilize only foundational Web standards; are simple to create and maintain for service providers; and can be discovered and utilized in a very intuitive way by biologist end-users. In addition, the SADI design patterns significantly improve the ability of software to automatically discover appropriate services based on user-needs, and automatically chain these into complex analytical workflows. We show that, when resources are exposed through SADI, data compliant with a given ontological model can be automatically gathered, or generated, from these distributed, non-coordinating resources - a behaviour we have not observed in any other Semantic system. Finally, we show that, using SADI, data dynamically generated from Web services can be explored in a manner very similar to data housed in static triple-stores, thus facilitating the intersection of Web services and Semantic Web technologies.Two Web technologies - Web services and the Semantic Web - hold the promise to achieve integration and interoperability among the currently disparate bioinformatics resources on the Web; however, this promise is not being widely achieved in practice. The causes of failure are varied, but often relate to the fundamental differences
%U http://www.jbiomedsem.com/content/2/1/8