Distinguishing between learning, growth and evolution

Abstract

Systems are not expected to stay the same over many versions; if there were no change at all there would be no improvement or continued satisfaction with a system. Lehman's laws of software evolution look at how a system changes over time. The current set of laws is now accepted as fundamental to the teaching and understanding of software engineering. This paper describes the terms "learning", "growth" and "evolution" from a biological perspective with a view to using and applying these ideas to an initial framework for the categorisation of software. The purpose is to classify software systems and thus understand more about their behaviour and characteristics.