19.2 Organization of the CWM Data Types

   The CWM DataTypes contains definitions of data types for the CORBA IDL language [CORBA], the SQL-99 language [SQL], and the Java programming language [Java]. Because they are M1 level entities, data type definitions for these languages are expressed in a tabular form that indicate the instances of M2 level CWM metaclasses that can be created in an appropriate CWM metadata store to define the M1 level data types. The data type definitions might then be used to create M1 level models appropriate for specific tools and software systems.

   The example M1 instances define only primitive data types; structured data types are not generally defined in these examples. (However, the CORBA IDL metamodel types required to define M1 structured types are provided as an example of how this might be done, if needed.) Data types that require no additional information to complete their definition, such as SQL’s INTEGER type, are completely defined. However, data types that are in some sense “parameterized,? such as SQL’s CHARACTER(n) and FIXED(p, s) data types, are incompletely defined because it is not practical to anticipate all possible parameter values! Tools that need to declare such parameterized data types should do so as they encounter them. The data type instances in this chapter define a few parameterized data types, where appropriate, as examples.

   As an example of appropriate usage of the CWM Foundation’s TypeMapping metamodel to indicate preferred and non-preferred mappings between the data types of different tools and software systems, the last section of the chapter contains some example mappings between Java and CORBA IDL and between Java and SQL-99.