Over the past few years I’ve been helping a lot of government clients design and implement architectures for information-sharing.  When it comes to information-sharing one of the most important aspects is the relationships in the data being shared because it’s those relationships that provide the context to help you understand and utilize the information better.  Within an enterprise, key data types are often dispersed across multiple systems so a lot of the things we did included implementing data services (i.e. SOAP/WSDL-based Web Services) on top of each of these systems to provide shared access to the underlying data.  While this approach does help share the systems’ underlying data with the rest of the enterprise, you still end up with these web services providing “islands of information”.  I.e. the relationships across the data being shared were not exposed and made easily traversable.  This is one of the reasons I prefer a REST-based approach for these more data-centric types of web services.  By following the REST principle of HATEOAS (Hypermedia As The Engine Of Application State), we explicitly expose these relationships across the shared data as links in the representations of the data that are returned from the data services and the links also make  it easy to traverse those relationships to access the related data.  This is in fact why information-sharing on the Web is so successful.  So by using a REST-based approach to create these data services, you can in effect re-create this architecture of the Web within the enterprise to enable effective information-sharing within the enterprise.