The preferred method for partitioning a logical data model is by creating submodels from packages. This approach has benefits including more flexibility and better partitioning of the logical data model in a team environment so that you can avoid having to merge changes. After you create a submodel from a package, you can open and access the sub-model through a shortcut in the main model, or you can open and work with a submodel directly. Containment is maintained between the main model and the submodels so that logical-to-physical data model transformation, impact analysis, and web publishing can be run transparently from the main model.