In support of the Ada 9X real-time facilities

Abstract

The Ada9x mapping has been criticised as being over ambitious and at odds with the “minimum change maximum benefit” AJPO dictum. Consequently there is a need to cut back the scope of the proposed changes. This report considers the proposed changes which affect the programming of real-time single processor and multiprocessor systems. We argue that protected records, select waiting, asynchronous transfer of control, nested asynchronous transfer of control are all essential core facilities , and if any of these are omitted the success of Ada 9x in addressing the needs of the real-time community would be in serious jeopardy. If language simplification are deemed necessary, we would advocate the removal of user-defined timers, finalisation (but not abort exit handlers), and protected record functions.