The toolkit provides two ways for an externalizer to obtain the full name of the class it is to instantiate: using toolkit definitions and using independent definitions. Toolkit definitions occur in the toolkit definition (btt.xml) file. This file has a keyed collection called tags that contains an inner keyed collection for each type of entity and these keyed collections define the mappings between the tag names and the implementation classes. Externalizers can also obtain the implementation class from the implClass attribute in the object's definition. The independent definitions provide greater flexibility by enabling implementation classes to change without requiring reinitializing the toolkit. The two ways of initializing objects can co-exist. The externalizer first determines whether tag contains the implClass attribute. If the attribute exists, the externalizer uses this implementation class even if the toolkit definition file contains a mapping for the tag. If the tag does not have an implClass attribute, the externalizer uses the toolkit definition file to obtain the implementation class. The externalizer sets the object's values using the values it obtains from one of the toolkit's generic definition files.

Processors can also group different entity definitions into a self-defined file. Self-defined processors enable applications to define new processors without having to change any of the generic type definition file and reinitializing the toolkit environment as a result of that change.