Software construction for mathematical applications

Abstract

This paper studies the problem of software construction for mathematical applications, and proposes a different structuring technique following the Object-Oriented paradigm. A short study about the “classic” functional structure used in this area is made, pointing out its qualities and drawbacks. An Object-Oriented analysis and design is performed, and from it is enumerated a set of requirements to impose on programming languages to use in implementation. At the end, the main advantages and disadvantages of this new structure are enumerated. To help to understand this work, one section is dedicated to an introduction to the usual software construction paradigms.