| Abstract: | When implementing large programs, it sometimes happens that the sheer size of the problem makes it difficult to apply recommended techniques for program designs. One example of this occurs when the size and complexity of a single data structure become great. It is naturally desirable that the details of the structure remain hidden from those modules which access it. For simple data structures this poses no problem since the data structure can be maintained by a separate data structure executive (effectively a cluster or class) which is the only module concerned with the physical details of the structure. However, for data structures which require very large executive procedures for maintenance and complex data declarations, it is necessary to partition the structure into substructures, each maintained by a separate module. If this is done in a straightforward manner, it will not be possible for other program modules to access properly parts of the structure unless that module can ‘see’ the various relationships of the substructures. Data metastructures are devices which, when properly employed, allow access to all parts of a partitioned data structure without the accessing module being aware of the partition. |
