Usage-based cost models for components

Cox (1996) argues that there is no economic incentive for building high-quality, reusable, software components. In his view, the mere notions of owning, selling and buying electronic property--such as software--form the key problem. The current pay-per-copy schemes for software distribution prevent easy experimentation with new software products and make it hard to establish a realistic price for each product that is based on its inherent characteristics such as, for instance, functionality, implementation effort, or lines of code. Instead, the distribution of software should be free and payment should be based on pay-per-use schemes. He proposes a new commercial infrastructure that opens up the potential for using large systems that are composed of many components, where each use of a component directly generates revenues for its owner/implementor. Since components can easily be replaced by better (and cheaper) ones, a competition between component manufacturers will emerge.

We agree with Cox's vision and observe that the software engineering approach we propose in this paper is compatible with it: pay-per-use schemes can easily be incorporated in the coordination architecture we have described. Issues to be addressed are security and confidentiality of the payment transactions and compatibility with emerging standards for electronic commerce.