While there are clear advantages to storing no more than a list of class names (see previous topic), this alone would not work well on an industrial scale. Diagrams would typically contain far to much detail to be of any practical value. AgileJ StructureViews applies sophisticated filtering rules to determine what to show, and how to show it. This collection of rules forms the Profile.
When a diagram is opened, the diagram editor reads the list of which classes and where to place them, calls the Eclipse Java Development Tooling plugin (the same plugin that drives the package explorer) to determine class members, then finally consults the Profile to determine how to display each member.
Here is the same diagram again, rendered differently according to different rules in the Profile. Class names are now unqualified, only private members are shown, the stereotype is hidden and lines a drawn differently.
