Is Software Engineering Really Engineering?

Abstract

The term ‘software Engineering’ emerged over two decades ago and is now commonly used to describe the process of software development in the fields of commerce and industry. In particular, the term is now taken to indicate a professional and rigorous approach to the development of software products. Within the united Kingdom this move to professionalism has recently been formally recognised by the acceptance by the Engineering Council of the British Computer Society as an Engineering Body, and the designation of its members as Chartered Engineers.

But do Software Engineers deserve the title? Are they really applying professional, rigorous approaches to the development of an engineering product? Are they utilising mathematical and statistical techniques to ensure quality? Are they working to standards? Indeed, do such standards exist? Are the products they produce really of high quality? In particular, is the software which is produced for use within traditional engineering disciplines (mechanical, electrical, civil) reliable? This paper will attempt to answer the above questions.

This paper will also focus upon the software “engineering” process in an attempt to draw parallels between software design and the design of traditionally engineered artifacts. Arguments in favour of the term “software engineering” will be made; in particular, the recent emergence of mathematical techniques for software specification, reliability prediction and measurement, and validation will be discussed. On the other side of the coin arguments ‘against’ will be made. Perhaps the strongest of these will focus upon the relative immaturity of the industry and the current lack of use of formal methods in industry and commerce.

The paper will conclude by assessing the current state of the art with regard to the “engineering” of engineering software and make recommendations as to future practices.