For a few years at the beginning of the 1990s, many people thought the applications software industry was on the brink of a far-reaching revolution. Objects and interlinked collections of objects called frameworks were going to replace packages. They would arrive down the wire, perhaps metered by the minute, as bundles of data and logic, and could be plugged into each other effortlessly to create new applications. No one was more interested in this revolution than IBM. With Apple Computer Inc and Hewlett-Packard Co, it created Taligent, a company initially set up to build an object-oriented operating system that would support applications built in this way. When Taligent was folded in 1995, it seemed the whole ideal was dead. But many of the Taligent engineers moved up to IBM’s AS/400 base in Rochester, Minnesota, and began working on a project called Shareable Frameworks, or SF a title that later became San Francisco (CI No 2,850). The goal is to produce frameworks, or components, that can be plugged together to create powerful, flexible enterprise business applications products that might be complex enough to rival the big ERP packages such as SAP R/3 or Baan. The project is now multi-platform, spanning Unix, Microsoft NT, AS/400 and mainframe systems. The fortunes of the San Francisco project were boosted by the adoption of Java, the fashionable object-oriented and internet-friendly programming environment. IBM does not plan to enter the applications business, but to work with partners that use San Francisco components to build applications. Joe Damasso, San Francisco project director of marketing, says IBM has three goals: to promote the use of Java, which it sees as important in electronic business; to promote the adoption of e- commerce using IBM systems; and to build a business providing the application components, for which IBM will charge a royalty.
Objects versus APIs
But why is this likely to appeal to the ERP software industry, where customers buy familiar packages, and where smaller software companies and services companies can make a living implementing or developing add-ons around the big packages? The main reason is that if true object-oriented technology is used rather than mere blocks of function, the software will be more flexible, easier to maintain, and more interoperable. For software partners, there is a big opportunity: they can offer customers quick, flexible, customized implementations. Damasso says IBM has finally got it right with San Francisco after earlier failures with object technology. [Before] components were too small. They were not prebuilt walls, rather a bag of cement, he says. The San Francisco components are not like the components of, for example, Baan or SAP R/3. Those are really blocks of function linked through an application programming interface (API). San Francisco objects have the technical characteristics of, for example, persistence and inheritance, which means they can be re-used and tightly linked without major programming. So far, 350 companies have signed up to the San Francisco licensing agreement, although this does not oblige them to develop products. Some influential software tools suppliers are also working with IBM on San Francisco, including Select Software Tools Ltd, Inprise Inc (aka Borland) and Rational Software Corp. The project is now delivering its first fruits. Objects and frameworks are being assembled into ‘towers’ which can be licensed by third parties. Among the towers that are complete or almost complete are: general ledger; accounts payable; accounts receivable; sales order management and warehouse management. Some partners are enthusiastic and have already begun projects with San Francisco- based technology. The first was Camelot, of Cambridge, Massachusetts, run by former-SAP AG executives, which is developing supply chain management software. It already has two or three corporate customers building big object based applications. Damasso’s view is that the ERP business is about to change fundamentally: At the moment there is exceptional growth at SAP, but that will slow and new start-ups based on SF will grow faster.
