The dBase User Group is dead, long live the dBase User Group. That was the main message from the dBase User Group meeting held in London recently. The community is split into three main camps: FoxPro devotees, Clipper fans and dBase IV diehards, all of which want to form separate user groups but will probably agree to come together under one umbrella user organisation. The meeting was very well attended – dBase users had come to hear what their future was, now that Borland International is in charge of the product’s strategy. However, the majority of people attending used FoxPro or Clipper and were more interested in the future of those products. Mark Daeche from Fox Software Inc was first on stage to outline the future of FoxPro. When it comes to futures, Daeche had the following warning to make: IBM bought a dream from Microsoft called OS/2, sold it to its developers and users and then Microsoft had a better dream Windows. It follows that the worst thing you can do is draw up a five-year plan and stick to it. That being said, FoxPro’s dream is to have the product available on Santa Cruz Operation Xenix, Windows and the Apple Computer Inc Macintosh by the third quarter of next year. All three versions are currently in beta test. As for OS/2, well that was dismissed as being as inviting as a bucket of spit. George Fletcher, managing director of Nantucket UK Ltd, was next to present, explaining the company’s two-pronged development strategy. The first of these offers users a gentle upgrade path from character-based Clipper 5.0 to full object-oriented capability. The second prong: Nantucket Future Technology is mostly being developed in Europe and will specifically address graphical user interfaces and operating systems. The next release of Clipper 5.0 will mix procedural and object code, supporting classes, instance variables and single inheritance – it will also support dynamic link libraries in Windows and OS/2. Developers will be able to upgrade applications written using Clipper in a modular fashion. Nantucket also plans to publish the Clipper 5.0 application programming interfaces, such as the replaceable database drivers. A series of different drivers will support SQL Server, NetWare SQL and so on. Nantucket’s other strategy – Future Technology – aka Aspen, is being developed simultaneously and will support user-interfaces and optimise the use of object-oriented programming techniques. This will facilitate multi-environment operations – that is, developers will be able to take an application that was written in one environment and move it across to another one.
Object layer
It will have database independence but will not be compatible with Clipper 5.0. However, for users of the object layer in 5.0, using this with Aspen will be similar as there will be shared syntax for the classes. After coffee, it was the turn of Borland’s Peter Rieks to explain his company’s take on the dBase market. Rieks said that both Paradox and dBase under MS-DOS will be maintained and versions of dBase will continue to be converted to Unix versions of the dBase language. Borland has built a new architecture under Windows and rewritten Paradox and dBase from scratch so that the Windows versions of both of these are not carrying heavy MS-DOS code. They share the same interactive engine, which can handle multiple file formats and referential integrity – the only difference from the developer’s point of view is the language used. All Borland products will support dynamic link libraries, dynamic data exchange and object linking and embedding. The object components in both character-based and graphical user interface-based environments will look almost the same to the end-user. The same dBase language will be used for developing both character-based and graphical interface-based applications. In other words, the Borland approach is to define common objects for the screen by using different programming methods. Borland provides ObjectVision, Paradox Application Language, dBase IV, C++, C or Pascal. The object layer is the differentiator
between the language methods, enabling them all to interoperate. Last to get up and speak was Dr Kathy Lang of Mayflower Consultants. Her biggest worry was the amount of hype Windows is getting. As she pointed out, Windows is difficult to implement in any rational, routine way – you can just as easily create a spaghetti interface with a graphical interface as with character-based text. But she was pleased to note that dBase does seem to have a future. – Katy Ring
