This page explains the ideaology behind the Workbench Shell used to tie code projects together.
One primary aspect of Workbench Shell is that everything works together in a project centric nature. The left pane is shows the current project that is being worked on. Currently, there is a fair amount of work left to be done on determining how to interact with the central project. The wiki will be updated as updates become available.
The Workbench Shell should be as loosely coupled as possible. The only tightly coupled addition like component is the Ontology Editor. All additions share the same instance of the Ontology Editor to maintain consistency. With the exception of the Ontology Editor, which is loaded at start up, all other additions use lazy loading. The additions.xml file is used to store necessary load information until the addition is activated for use. This keeps in use memory and start up times to a minimum. For instructions on how to add an addition see Workbench Additions.
Note the intention is that each addition can be run standalone and completely independently of the WorkbenchShell. The WorkbenchShell merely provides a convenient way to use tools together under a central hub. Using Eclipse terminology, each addition is a view because only one instance is allowed of a single addition at a time. Design your additions accordingly. This architecture may change at some point in the future, but for now this is sufficient for our needs. Note that this architecture still allows for editor type functionality, so long as it is controlled by the addition (think Ontology Editor).