In introducing its own, and yet another, application development toolset, SAP AG has taken some pages from Microsoft Corp’s playbook. At the introduction of ABAP/4 in Karlsruhe, Germany, last month, SAP co-founder and vice-chairman Hasso Plattner was asked why he would add to market chaos with his own product. With refreshing honesty about his company’s market dominance, Plattner said, I’ve learned some things from Bill Gates: this is capitalism. I’m not against CORBA, but these initiatives and the Open Software Foundation and Common Desktop Environment – they are all relatively slow. OLE is here. It’s not perfect, but it’s a start. Why not use it instead of discussing why it’s not perfect? There are already many products out there supporting OLE, while CORBA is the subject of academic discussions. We don’t want to wait for someone to define something that we might be able to use. SAP has just unbundled ABAP/4 from its suite of business applications. It runs on all Unix and Windows NT systems, supports all widely-used graphical user interfaces and accesses databases via SAP’s Open SQL or calls to the vendor’s native SQL derivative. It has existing interfaces to applications such as Microsoft’s Visual Basic, Excel and Access. The product will be used first by consultants and partner developers, he said. It is priced at from $50,000 for a five-user developer licence, including the R/3 BASIS middleware component. Each additional user costs $6,750. We feel comfortable enough with R/3, that it is a solid enough system that we can spend a significant amount of resources to talk to the customer at the development level, Plattner said. In fact, as he indicated in a presentation to hundreds of German users earlier that morning, SAP will have spent $640m on R/3 by the end of next year, The most expensive software product developed by anyone, ever. SAP says R/3 has 2,000 installations among 1,400 customers. So far, he said, the company has had little feedback on release 2.2, which was shipped in September. Release 2.2B will be in general release very soon, with some major enhancements for manufacturing and distribution gleaned from sites in Germany and the US, he said. Contrary to indications from its rival Oracle Corp, Plattner said he feels no vulnerability at all in the process manufacturing sector. Release 3.0 of R/3 will include a Product Planning and Control for the Process Industry module that covers lab and process control equipment and recipe handling, among other functions, he said.

Starting to copy

Furthermore, he said, SAP’s five-year investment in data and process models is beginning to pay off. Competitors are starting to copy them. We doubted the prudence of publishing them, but so far the competitive products only look the same. Once you get past a few screens, you see that they lack data sophistication, he said. Walldorf is not Hollywood, and SAP will not build an R/4, Plattner said. R/3 has been built so that enhancements can be added on the fly. We are starting to re-engineer the system, however, and the first two of those projects will be part of release 3.0, he said. That release, which is scheduled to ship in the second quarter of 1995, will be enabled as an Object Linking & Embedding automation server and client, will have more integration with desktop software and less downtime for maintenance upgrades. Plattner made reference to SAP’s ‘Project Heidelberg,’ which is a plug-and-play version of R/3 for Windows NT based on Microsoft’s Windows95 and Oracle’s workgroup server. It will have some industry-specific functionality and be convertible into standard R/3. The software will be used first by US firms with sales of $50m or less, he said. The chief executive also outlined SAP’s broad development strategies and principles: reliance on Remote Function Call for integration; Object Linking & Embedding and SAP-Application Link Embedding are key initiatives; co-operation between internal and external workflow; redefinition of distributed system architecture; the need for a homogeneous set of applica

tions for distributed databases; and diversification of local business practices making sharing the same data impossible. It’s a mistake to believe that all possible forms of business practices can be reflected in a single physical system. Distribution [of applications] will be possible using the technology of object link exchange. The distributed database is a myth. Distributed database technology can be used for propagation of shared meta-data, such as repository objects, Plattner concluded.