We suggest that there is a lack of empirical evidence for claims that object-oriented software has an advantage in terms of software maintenance. We use a model of software maintenance that suggests that the maintenance task and programming language have an impact on the maintenance process and, in turn, use the maintenance process to explore the impact of the programming language on the maintenance process. A field study design, using students, examined the perceived difference in process employed by students using an object-oriented language and those using a third-generation language. We found a difference in the perceived process used by the two groups of students. Students using the object-oriented language perceived that they performed more planning and diagnosis activities during the maintenance process and students using a third-generation language perceived that they performed more knowledge building activities. Copyright © 2007 John Wiley & Sons, Ltd.