Die-hard Cobol programmers are unlikely to convert to fourth generation languages until they contain commands equivalent in power to more traditional third generation languages, according to Mark Rogers of the UK software firm MF Systems Ltd, based in London’s Kensington. According to Rogers, talking at the recent PC User Conference at Olympia, many fourth generation products have been self-defeating, helping developers with the easy 80% of programming tasks, but offering no help to the more complex 20%. It’s almost regarded as acceptable that 4GLs should provide links to languages like C for the complex pieces of the program, said Rogers. But as soon as you do that, you forfeit most of the benefits you gained from using a 4GL in the first place; you lose development speed, ease of maintenance and portability. Problems of efficiency, a major concern in the early days of fourth generation languages, is in most cases no longer an issue, with efficient coding overcoming performance problems and virtual memory techniques optimising the use of memory. And new database techniques such as variable length fields use far less disk space than traditional systems, and are faster to use. Apart from development speed, Rogers pointed to maintenance of programs as a key benefit of switching to fourth generation langauges. More than 75% of a programmer’s time is spent amending existing software, rather than developing new programs. By using a data dictionary, fourth generation techniques make it easier to amend systems, and to manage projects where there are different versions of software. But future products should concentrate on increased sophistication rather than simply aiming to save time and money. The real worth of using a 4GL is to write better, more sophisticated programs which could not be written before, suggested Rogers. I feel this objective has been somewhat neglected.
