We describe the requirements -- scientific, technical and economic -- of scientists and engineers who are developing new meteorological algorithms and products. We argue that these requirements call for quick and ready access to multiple, disparate sources of data (including the outputs of other algorithms) in a uniform manner. We also argue that developer convenience dictates that it be possible to develop and test algorithms using data from an operational system without impacting that system. For efficiency, algorithm developers should be able to run multiple instances of any algorithm and to compare the performances of the different instances. It should be possible to easily visualize the intermediate and final outputs of the algorithms alongside the original data and the outputs of other algorithms. Application Programming Interfaces (APIs) geared towards algorithm development should provide all of these features to create a robust community of meteorological algorithm developers.

In this paper, we describe an API that meets these developer needs. Other papers in this session describe new algorithms that have been developed using this API, called the Warning Decision Support System -- Integrated Information (WDSS-II). These algorithms illustrate different facets, both of the requirements of new algorithms and of the capabilities of the API.