Article Preview
Top1. Introduction
The software maintenance or evolution is an essential step in the software life cycle. Several works have, for many years, highlighted the importance of maintenance and/or evolution of the software and have established several taxonomies (Gasmallah et al., 2016). The oldest have allowed to consider three kinds of maintenance (Swanson, 1976): corrective maintenance leading to remove the residual errors or faults; adaptive maintenance that consists of adapting the software to changes affecting both its technical or managerial environments like the evolution of the deployment infrastructures or the change of some regulation policies, etc. The perfective maintenance includes changes intended to improve the software by introducing new features or improving some quality criteria, etc. In (Chapin, 2000), the author introduces a new maintenance type called preventive maintenance as the changes leading to make the software more able to evolve and then to change. In other words, the changes induced by this last kind of maintenance makes it possible to improve the maintainability criterion of the software without affecting the functionalities or performance of such a software.
In this work, we do not intentionally make any difference between maintenance and evolution, although these two concepts have been subjects to numerous comparisons. Maintenance is often seen as an engineering process in which academic research has delineated and characterized the stages, the activities to be carried out, the actors and resources to be implemented, etc. In other words, the concept of maintenance is the result of a kind of morphism between the industrial world and the software development activity. The term software evolution is the result of works (Lehman & Ramil, 2002.) aimed at understanding the phenomenon of change affecting the software artifacts. For example, Lehmann's laws of evolution attempt to explain how the software evolves during many years until becoming obsolete. These laws are the result of empirical studies conducted over several years on software used on the real world. In the same way, Benett and Rajlich drew up (Benett & Rajlich, 2000) a software life cycle designed not as a means to develop the software but as an explanation of how software is produced, evolved, maintained, and eventually removed from the information system.
As far as we are concerned, we indifferently use the terms maintenance or evolution. What we are particularly interested in, is the notion of the software change and more specifically the Change Impact Analysis (Abdi et al., 2009; Sun et al., 2012). It is clear that one of the major concerns of the software development stakeholders is to best control the change process and then the software maintenance/evolution one. In fact, an uncontrolled change can quickly lead to a slippage in terms of project’s cost and time. During more than four decades, many works have addressed this problem making the maintenance the most important cost factor of all the software development process (Folmer & Bosch, 2008; Gupta et al., 2008). In addition, according to ISO 250010 Model (ISO, 2011) (José et al., 2014), maintainability or scalability is one of the main components of the software quality.