The Avatars of a Package for Calendars in Ada

ions of real life, as opposed to concrete objects. Our specification has shown the ability of the language to follow the principles of software engineering. In particular 1. Encapsulate everything that does not exactly coincide with its implementation. 2. Hide everything that is not required to be visible. 3. Report possible failures and make it possible for the user to perform the appropriate action. 4. Distinguish between inherently impossible cases and implementation limits. It seems also useful to recall the following style rule: 5. Declare exceptions in a non-generic package common to the whole of each subsystem; in Ada-95, this package could be the root of a hierarchy of components that makes up the subsystem. Part of the Ada details can seem to be very pointed. The robustness improvement, compared to other languages, has, of course, a non-zero cost. There is no obligation to choose such a highly robust specification, you may prefer an intermediate solution, according to your specific needs and quality requirements. Part of the Ada details can seem to be very complex to implement. They are in fact basic techniques and, again, this complexity is part of the cost of quality. The more they are used, the less complex they will become. In any case, the contents of this paper can be used as a source of ideas for training exercises. For instance: (a) implement arrays indexed by dates, or by any conceptually discrete type, and instantiate it with dates (b) specify types for physical measures so that for example metres divided by seconds yield metres-per-second (c) specify multi-currency computations (d) specify a read-only view of existing object-class packages. Many parts of this program are useful. None are totally useless, exotic or artificial. Their design from the apparent one-package form to the final multi-component library is an illustration of what could reasonably be called real-based programming, which tries to keep the design as close as possible to the modelled reality rather than to satisfy programming language requirements. Considering this, it is not bound to a particular programming language. Programming languages can be well suited (or not) for easily implementing real-based designs which requires a high level of versatility from the language. This paper has shown that Ada is suitable for this, whereas some other languages lack some essential features (namely, all those where the operators are implicitly declared and impossible to hide). Nothing has been said about the possible existence of other languages with the required features. From another point of view, this paper can be taken as a plea in favour of such languages. WHERE CAN THESE CALENDARS BE FOUND? The Gregorian, Julian, Muslim, Hebrew and French Revolutionary calendars have been built, some of them as student projects, together with their test programs. However, they are beta versions and code review must be done before they are made available. Thanks in advance for any comment. 427 the avatars of a package for calendars Ada-France have designed a repository, called CAFE (a French word that means both coffee (the drink) and café (the shop), with the original French meaning of the acronym ‘Composants Ada pour la Formation et l’Enseignement’, but with the possible English meaning ‘Components in Ada For Education’), for Ada components with specific design or robustness, suitable for Education. The calendar components are ftp-able from ftp.unilim.fr, and freely usable under the ‘General Public Licence’ of the Free Software Foundation.