Distributed application designed using MASCOT and implemented in Ada

Abstract

The prime objectives of the study are to demonstrate the feasibility or otherwise of the current MASCOT 3/Ada technology to produce a typical Command & Control System and to provide a benchmark facility for Ada compilers and run-time environments.

We have satisfied the requirement to assess the feasibility of the current MASCOT 3/Ada technology by producing a substantial body of Ada code and considering the problems in scaling up to a large system. It is our opinion, that with the minor modifications to MASCOT 3 and Ada as outlined below and with the automated support of a MASCOT 3 sensitive CASE tool, this technology will be applicable to the development of large, real-time systems.

There are still inadequacies in many Ada tools — especially debuggers. Limitations in tools can cause serious difficulties if there are direct conflicts of requirement versus capability. Problems encountered are often tool limitations rather than inadequacies of the Ada language itself. There will always be unsatisfactory ways of using particular language features and unfortunately tools often perform badly when these ways were originally unanticipated by the vendors.

The techniques and tools used on this project proved adequate, but for larger systems program generators must be able to produce incremental code changes: regeneration of the whole system is simply too much to do for every application-level change.Overall, the project was able to draw together the various strands of earlier work and did succeed in its aims and demonstrated that:

• 1 Mascot3 can successfully map to Ada code for a project of non-trivial size,

• 2 Ada that is genuinely portable between different compilers can be written,

• 3 Heterogeneous processors can operate together even using different compilers.