Software Maintenance Research Research Articles

ID-correspondence: a measure for detecting evolutionary coupling

Evolutionary coupling is a well investigated phenomenon in software maintenance research and practice. Association rules and two related measures, support and confidence, have been used to identify evolutionary coupling among program entities. However, these measures only emphasize the co-change (i.e., changing together) frequency of entities and cannot determine whether the entities co-evolved by experiencing related changes. Consequently, the approach reports false positives and fails to detect evolutionary coupling among infrequently co-changed entities. We propose a new measure, identifier correspondence (id-correspondence), that quantifies the extent to which changes that occurred to the co-changed entities are related based on identifier similarity. Identifiers are the names given to different program entities such as variables, methods, classes, packages, interfaces, structures, unions etc. We use Dice-Sorensen co-efficient for measuring lexical similarity between the identifiers involved in the changed lines of the co-changed entities. Our investigation on thousands of revisions from nine subject systems covering three programming languages shows that id-correspondence can considerably improve the detection accuracy of evolutionary coupling. It outperforms the existing state-of-the-art evolutionary coupling based techniques with significantly higher recall and F-score in predicting future co-change candidates.

Software maintenance research that is empirically valid and useful in practice

Abstract For successful software system, maintenance efforts dominate initial development costs by far. However, research and practice still place most of the attention on the initial development phase. In consequence, companies suffer mission-critical systems that are extremely hard to maintain. To better support practitioners in writing maintainable software, we both need an empirically sound understanding of software properties that help or hinder maintenance and tools that are useful in practice. This requires work that includes both research and practitioner perspectives. In this article, I outline my work towards this goal.

An exploratory analysis of the state of software maintenance research

PurposeResources allocated to software maintenance constitutes a major portion of the total life cycle cost of a system. The enormous effect that this can have on an organization's ability to react to dynamic environments has been the primary motivation for researchers to study issues related to software maintenance. The purpose of this paper is to take stock of the research conducted in this area in order to identify the intellectual trails embodied in the coherent body of knowledge on software maintenance.Design/methodology/approachAn author co‐citation analysis (ACA) involving authors who have made seminal contributions to the field of software maintenance was performed. The data for the study were obtained from the Science Citation Index (SCI) and the Social Sciences Citation Index (SSCI).FindingsThe results indicate that most of the software maintenance research has focused on eight areas: Program logic characteristics, Quality of processes/metrics, Effort and productivity issues, Cognitive issues in repair maintenance, Organizational Issues: Strategies for software evolution/maintenance, Object‐oriented (OO) maintenance, Domain specific language issues and Program construction and design. Research limitations/implications – Some of the limitations of this study include: exclusion of data after 2003, giving equal weight for all citations, and implicit assumption that a relationship exists between the citing and cited documents.Practical implicationsThe extension of software maintenance research in the areas suggested in this study may lead to new innovations for practice.Originality/valueThe paper not only introduces new methods for meta analysis, it also suggests that opportunities abound for extending the frontiers of software maintenance research, particularly in the context of contemporary software development approaches.

An ERP-client benefit-oriented maintenance taxonomy

The worldwide installed base of enterprise resource planning (ERP) systems has increased rapidly over the past 10 years now comprising tens of thousands of installations in large- and medium-sized organizations and millions of licensed users. Similar to traditional information systems (IS), ERP systems must be maintained and upgraded. It is therefore not surprising that ERP maintenance activities have become the largest budget provision in the IS departments of many ERP-using organizations. Yet, there has been limited study of ERP maintenance activities. Are they simply instances of traditional software maintenance activities to which traditional software maintenance research findings can be generalized? Or are they fundamentally different, such that new research, specific to ERP maintenance, is required to help alleviate the ERP maintenance burden? This paper reports a case study of a large organization that implemented ERP (an SAP system) more than three years ago. From the case study and data collected, we observe the following distinctions of ERP maintenance: (1) the ERP-using organization, in addition to addressing internally originated change-requests, also implements maintenance introduced by the vendor; (2) requests for user-support concerning the ERP system behavior, function and training constitute a main part of ERP maintenance activity; and (3) similar to the in-house software environment, enhancement is the major maintenance activity in the ERP environment, encompassing almost 64% of the total change-request effort. In light of these and other findings, we ultimately: (1) propose a clear and precise definition of ERP maintenance; (2) conclude that ERP maintenance cannot be sufficiently described by existing software maintenance taxonomies; and (3) propose a benefits-oriented taxonomy, that better represents ERP maintenance activities. Three salient dimensions (for characterizing requests) incorporated in the proposed ERP maintenance taxonomy are: (1) who is the maintenance source? (2) why is it important to service the request? and (3) what––whether there is any impact of implementing the request on the installed module(s)?

Supporting software maintenance with software engineering tools: A Computed task–technology fit analysis

Management has turned to software engineering tools designed to support software maintenance as a potential solution to maintenance productivity and quality problems. Once adopted by an organisation, however, these tools are often not used. A research model, based on a task–technology fit (TTF) model, is developed to explain the factors which lead to the use of the software maintenance support tools. Our model, which examines the nature of the fit between software tool functionality and maintenance task demands, consists of a TTF model augmented with a model of the maintenance task and a model of software maintenance tool functionality. Such an augmented model is necessary for moving beyond isolated exploratory studies of maintenance and for building on the existing research in software development and software tool utilization. Using this model, fit between maintenance task and technology characteristics is computed for two dimensions of fit derived from the task and technology models. Tests of hypotheses derived from the model demonstrate that task–technology fit, computed using methods for computing strategic fit, is associated with increased tool utilization. Our findings provide direction for the development of better maintenance support tools. This research extends the application and usefulness of TTF models and maintenance task and technology models so that future software maintenance research can build on tested models.

Characteristics of empirical software maintenance studies: 1980–1989

AbstractThis study examines empirical software maintenance research during the decade of the 1980s and categorizes the research according to placement, origin, and nature of research to determine what has been done in the past. To aid in the investigation, an exhaustive electronic keyword search was made of abstracts found in Computer Database Abstracts and INSPEC's Computer and Control Abstracts. Once candidate articles were identified, their full abstracts were scrutinized to determine whether the entire article should be reviewed. Based on a review of the full articles, summary statistics and research trends were established for software maintenance research published during the decade of the 1980s. The discussion that follows is partially based on an interpretation of the findings by an expert panel of information systems professionals actively working in the software maintenance area or engaged in management activities associated with software maintenance. Conclusions concerning the form and direction of future software maintenance publications are then advanced.