Using large families for handling priority requests

With ordinary task entries, requests are handled strictly in the order in which they arrive at the called task. To circumvent this rule, for prioritised requests, the Ada language provides entry families. Unfortunately the programming of a reliable algorithm for a large number of priority levels is non-trivial. This paper reviews the "solutions" usually given and derives an acceptable, if somewhat complex, algorithm. Suggestions for improvements in the facilities Ada provides are also presented.