Contemporary Software Modernization: Perspectives and Challenges to Deal with Legacy Systems

TL;DR

This paper reviews the evolving field of software modernization, emphasizing the need for flexible, up-to-date approaches to update legacy systems and proposing a research agenda with key challenges.

Contribution

It highlights the limitations of existing modernization approaches and advocates for a comprehensive research agenda to advance the field.

Findings

Current approaches are too scenario-specific and inflexible.

Many modernization methods are outdated and not aligned with modern development.

The paper proposes 10 research challenges to guide future studies.

Abstract

Software modernization is an inherent activity of software engineering, as technology advances and systems inevitably become outdated. The term "software modernization" emerged as a research topic in the early 2000s, with a differentiation from traditional software evolution. Studies on this topic became popular due to new programming paradigms, technologies, and architectural styles. Given the pervasive nature of software today, modernizing legacy systems is paramount to provide users with competitive and innovative products and services. Despite the large amount of work available in the literature, there are significant limitations: (i) proposed approaches are strictly specific to one scenario or technology, lacking flexibility; (ii) most of the proposed approaches are not aligned with the current modern software development scenario; and (iii) due to a myriad of proposed modernization approaches, practitioners may be misguided on how to modernize legacies. In this work, our goal is to call attention to the need for advances in research and practices toward a well-defined software modernization domain. The focus is on enabling organizations to preserve the knowledge represented in legacy systems while taking advantages of disruptive and emerging technologies. Based on this goal, we put the different perspectives of software modernization in the context of contemporary software development. We also present a research agenda with 10 challenges to motivate new studies.