Ada 9X: a technical summary

Reviewer: Brian Lawrence Meek

The revision of the original Ada language dubbed “Ada 9X” has (as of October 1993) reached the stage where a draft revised standard and a rationale [1,2] have been submitted for wider review within the standards bodies. That accomplished , the draft (if approved) or a modified version will go for balloting as a draft international standard. This means that the design team believes its technical work to be largely complete, and that any technical changes between now and final approval are likely to be minor. Time is therefore running out for those who would seek to influence the final form. Now is also a good time for those who might be interested in using the revised Ada to start looking at what more it will offer compared to what we must now start calling “Ada 83.” This paper provides an overview of some principal features of the revision, and hence forms a good place to start. Some familiarity with Ada 83 is of course assumed. Taft regards the principal changes as being in three areas. For object orientation, the Ada datatype model is extended to support full polymorphism; in line with Ada's strong datatyping tradition, exact datatype matching is the default, but the draft standard provides for classwide matching if explicitly desired. For programming-in-the-large, separate compilation facilities are enhanced, standard mechanisms are introduced for partitioning large applications (while preserving strong datatyping), and the library namespace becomes hierarchical rather than flat. For real-time and parallel processing, support for asynchronous communications is added to the existing synchronous mechanism based on the rendezvous mechanism. The author clearly and concisely reviews the Ada 9X developments in all three of these areas, and provides some code extracts as illustration. I recommend this paper to any Ada user as a quick and easy introduction to the Ada 9X extensions. The only possible caveat is that some writers would have included a fourth principal area—interfacing to non-Ada systems and other languages—so if that is a particular interest of yours, you will need to look elsewhere. The paper will also repay attention by those who may previously for various reasons have steered clear of Ada, particularly people of an object-oriented persuasion, of whatever flavor. If you expect to want to develop large industrial-strength applications using object orientation, Ada 9X may be what you are looking for.