7.1 Implementation Issues

SIESTA is based on a LISP interpreter called VLISP [4,35]. There are several arguments for the usage of VLISP and some against its usage. The strong points are:

• The LISP language itself provides inherent capabilities for manipulation of data sets with variable size and heterogeneous data through its lists. This is comparable to functionality emerging lately with the standard template library of C++ (although it may be slower in LISP).

• VLISP offers a highly developed object system (see also Appendix A) which encourages modularity, simplifies maintenance, and speeds up development.

• Dynamic invocation of objects of arbitrary classes is possible with VLISP.

• VLISP offers a binding to the MOTIF toolkit [17,31]. That means that the most elaborate and widespread toolkit available on UNIX platforms is available from within SIESTA.

• LISP's inherent parser can be used for various purposes.
• The JAVA [10,21] language was not an industry standard when the development of SIESTA started. Moreover the GUI features provided by JAVA through the Abstract Windowing Toolkit (AWT) lack a lot of functionality which is offered by MOTIF. And its successor, the SWING toolkit, was not available at all when the development of SIESTA was started.

Arguments against the usage of VLISP are:

• LISP is not very popular.
• LISP is not an industry standard.
• Programs written in LISP are slower compared to programs which have been written in languages like C or C++.

However, since the advantages of using VLISP were dominant, it was selected to be the language of choice. It should be stressed that all the concepts that are presented in this work, are in no way related to VLISP or specific features of VLISP. To some extent, however, VLISP's object oriented features, which are also available in some other languages, simplified the implementation of the anticipated functionality.