The Eclipse Resource Management defines a set of APIs for managing file-system-based workspace resources. It provides the foundation for various other services of the Eclipse platform, ranging from editor support and resource markers to Team support and Mylyn. With the Eclipse File System (EFS) API, it is now possible to plug-in alternative file system implementations e.g. to access network file systems or other content storages. The EFS is a great thing, but does the whole world look like a huge distributed hard drive with strictly hierarchic file/folder structure? There is a lot of resource-oriented content offered over network via a variety of protocols ranging from REST-based HTTP over WS-Transfer/WS-Resource to proprietary/legacy ones. Very often the resource-based content is offered in a way that doesn't really fit to a hierarchic access pattern. There are also addressing schemata for resources, like hierarchic URIs and UUIDs that can’t be directly mapped to a file path. Looks like there is a big gap between the EFS-based access to file system content and the RESTful or REST-like universe of resources. Should we give up Eclipse Resource management and look for something else? Not necessarily.
This presentation briefly describes a proposal for an add-on to the Eclipse Resource Management that tries to bridge the gap and augment the Resource Management APIs with additional standardized APIs to work with RESTful or REST-like resources. The implementation is based on EFS and requires no changes to Eclipse Resource Management. The add-on introduces a new SPI to be implemented by adapters that handle access via their specific communication protocol or proprietary client library.