At the Micro Focus second European User Conference held in London recently, an enthusiastic audience of data processing managers squeezed into a meeting room to hear the latest about object-oriented Cobol from Jerome Garfunkel and Raymond Obin. This is obviously a very magnetic topic for Cobollers and legitimately so, since as Garfunkel puts it object-oriented Cobol is off and running. There are two alternative ways of adding object-orientation to a language – it can be done consistently so that the object-oriented words and classes appear the same as regular Cobol, or new object-oriented syntax can be added on top of the Cobol language. The former approach is being adopted, however, because the Codasyl Cobol committee decided that if the other, orthogonal, approach was taken there was really no point in using Cobol. One of the most interesting possibilities of object-oriented Cobol is the potential to re-use all those billions of lines of Cobol already in existence. Indeed, it has been estimated that building object-orientation into Cobol could result in some large Cobol installations cutting down new program development by as much as 70%. To this end the Cobol committee has laid down as a goal that old code should be encapsulated in an object wrapper so that the current huge inventory of Cobol programs will be able to work with the newly created and designed object-oriented Cobol programs. Garfunkel says he is hoping that such encapsulation will be implemented as an automatic process. He, personally, is quite enthusiastic about object-oriented Cobol, but has a couple of reservations: firstly, he thinks that issues such as software re-usability could have been addressed without object-oriented methods, but that there simply has not been the political will and he is not fully convinced that the will exists now. Secondly, he knows that programmers are not really aware how radical a change object-oriented techniques offer particularly in the cataloguing and management of the objects. However, vendors such as Micro Focus really do believe in the new technology and its attractions are obvious. It means that programmers can write applications that mimic the real world, since with object-oriented methods they can invent new Cobol syntax to do so. Programmers will be able to create new verbs indeed, Garfunkel believes that a whole new job definition will come into being – the Object Designers whose job within a company will be to invent new verbs and new data types. He also thinks that a whole new industry will spring up as young companies will create libraries of objects and sell them as off-the-shelf packages. In his opinion the first object-oriented Cobol compiler will appear by the end of next year. One possible, and to Garfunkel’s mind, very welcome, side effect of this move into object-orientation could be that Cobol becomes acceptable within universities, which means that programmers will be able to leave academe and have a higher chance of walking straight into a job. However, if there are massive cultural problems for programmers getting to grips with object-oriented Cobol, so too are there problems for the managers of those programmers. For a start they will need programming tools: something to edit multiple classes with, a class locator to find the relevant class to re-use, a class library to interface to the locator and so on. And then they must also help to develop the appropriate culture for object-oriented programming. As Adele Goldberg puts it: You get the behaviour you reward, so reward the behaviour you want. With object-oriented Cobol this will mean above all rewarding the programmers who make use of other people’s code and who curb their instinct to meddle with a program and rely on system logic instead. After all, with object-oriented programming you need only know that something is implemented and how to use it. For the difference for programmers between the old and the new Cobol will be as great as that between a car mechanic and a car owner that carries out regular car maintenance. – Katy Ring
