The US Department of Commerce’s National Institute of Standards & Technology has awarded a $19.4m five-year grant to a group of computer companies and universities that are working to create software development approaches in semantic composition and component-based software. The grant has been made under the Advanced Technology Program, and it has been one of the largest the institute has awarded so far during fiscal 1995; the programme, however, has just lost $90m from its $431m budget. The money was rescinded by the US government and this has meant that certain competitions for research funds have had to be delayed until the next financial year, which starts in October. But the institute does not believe that the cuts will adversely affect the component-based software programme; it wants the group to develop software components that people could buy off the shelf and configure into their own customised applications. IBM Corp, Hewlett-Packard Co, Apple Computer Inc, Motorola Inc, Baxter Inte rnational Inc, Taligent Inc, Kestrel Development Corp, Stanford University, SRI International and the University of Southern California are all members of the joint venture. Dr David Fisher, manager of the Advanced Technology Program, said the idea is to create a commerce in components, a move seen as crucial by the institute to the US’s continued dominance of the world’s software markets. Companies and organisations receiving the money will develop and market small, broadly useful software components described by formal specifications that characterise the logical and functional requirements and characteristics of components. Fisher describes the notion of commerce in components as akin to the trade in something like screws: people know what they look like, know they can come in different sizes and can produce them without needing to know to what purpose they will be put. There is no software assembly business because there is no component sales business, he said, but any other market has this and the ensuing benefits. The barrier in the software business is purely technical. The institute reckons that off-the-shelf software for personal computers makes up about 15% of the market. The remaining 85% is large custom-made programs often cos ting millions of dollars to create. This is a situation the institute believes is unwieldy and which eventually has to be challenged. It wants to ensure that the US is at the forefront of any technical advances in this direction. But components should not be confused with object technology: objects are system-specific technology, whereas the components to be developed would work on any system and are much more fine-grained: each component describes one particular function within an application whereas each object could be made up of numerous components. And the institute would like the joint venture to design development tools that could automate the ‘weaving of the components’ into applications. This is the semantic composition part of the program me. Fisher said the group is being asked to develop formal methods or, as it is known in Europe, non-procedural language methods, that would automatically systhesise programs as required.
Off the shelf
The theory is that if both components and development tools were available then users could buy all they needed to, off the shelf and automatically configure an application that did everything the user wanted. Software companies could then concentrate on meeting application needs, rather than the mechanics of the software development process. We would then have a viable marketplace for software, Fisher said, adding that the advent of such a system would revolutionise the world of software manufacturing. The joint venture has attracted some big names but not the biggest, Microsoft Corp, which, of course, might be reluctant to see users buying components to build their own applications. But Fisher says that Microsoft would probably not have qualified for funding as the programme will not support projects that companies could fund t
hrough free market sources. Those taking part had to define an approach to develop scalable, automated semantic-based software, focusing on testing, automatic generation of code to link software components together, and the creation of custom languages that automatically generate software components. Motorola, not known for its software technologies but keen to get into the field, will provide the strategic direction to ensure that efforts support commercialisation of the software technologies and to develop proof-of-concepts or demonstrations of the technologies.
