Compact database sizes, random access, and short read and write times are requirements that cannot be fulfilled with an ASCII format. A binary representation of the interchange format is required to meet these goals. This binary representation functions as a kind of simulation database which stores simulation problem descriptions, histories and results.
Early implementations of binary interchange formats, like the DAMSEL system from CNS/CNET [Corb88], feature two-dimensional geometries and simple data structures for easy usage by existing simulators. The DAMSEL format is used as an interchange format in the joint European Community STORM project [Jone93][Lore93]. A more advanced implementation designed for TCAD environments is the CDB/HCDB from CMU [Walk93a], which functions as a binary interchange format in the PREDITOR TCAD framework [Walk93b].
Many TCAD systems use PIF in a binary form as an interchange format. All of those have the possibility to translate the PIF file into a textual representation as described above. The base of the VISTA data level is also built upon a binary PIF implementation see Section 3.4.
The EASE (Electronic Applications Simulator Environment) system of ABB uses a PIF derivative called UDIF as an interchange format [Stra90]. The BPIF (Binary PIF) implementation from UC Berkeley [Wong91] functions as the data level for the PROSE environment [Wong92][Neur93]. It uses the OCT database system [Harr90][Barn92] as a base upon which the PIF binary data structures are modelled.
Another data level built on PIF featuring object-orientedness is the PIF/GESTALT system from MIT [Boni91] used in the CAFE (Computer Aided Fabrication Environment) system [McIl90]. This implementation is also based on the OCT database system.