Back to search

Πανεπιστήμιο Μακεδονίας

Managing technical debt in modern software applications

Abstract

dc:description

Technical Debt (TD) is a metaphor which is used to describe the software inefficiencies that are introduced in the code base as “the price to pay” for the quicker delivery of the product. In recent years TD has attracted a lot of attention from academia and industry, probably due to the ability of the metaphor to bridge the gap between business and technical stakeholders. Driven by this, we have observed an increased attention to the process of TD management, with the introduction of tools or specific TD activities from the quantification all the way to its repayment. Even though the TD metaphor was introduced more than three decades ago, there are still a lot of aspects that are underexplored or need to be continuously examined following the advanced in the software development process: •it is well-known that to reduce TD a software engineer needs to perform refactorings (the most prominent TD repayment method); however, this activity alone is not able to prevent TD accumulation. As a result, there is a need to explore which activities in the development process affect TD accumulation, and if these activities can be performed at management or engineering level. To this end, developers and managers need to be aware of the effect that each activity can have on TD, and what actions they can take to reduce TD. •given the fact that the software development practices rapidly evolve and change continuously, TD management approaches need to be updated. More specifically, specific modern software development methods (e.g., service-oriented architectures) might require a different approach for TD quantification and management approaches. For example, the organization of SOA applications is different com-pared to traditional object-oriented applications, yielding for different level of analysis. This highlights the need for novel methods that would be more accurate for emerging domains. To this end, the purpose of this dissertation is to advance the state of the art in the fields of TD management, TD repayment, and TD quantification keeping in mind and adapting to the new development practices that arise. Following the first line of research, in this PhD project we attempt to understand the rea-sons of TD accumulation. Initially, we categorized quality improvements activities into two classes: (a) micro-management: interventions applied at a fine-grained level (e.g., at a class level by applying a refactoring); and (b) macro-management: interventions applied at a large-scale (e.g., at project level, by using a new framework or imposing a quality gate); and next we examine the relation of these activities to TD accumulation. Based on the results, we confirm that micro-management parameters are associated with the probability of a class having high TD, but the results can be further improved by controlling some project-level parameters as well. Moreover, we moved deeper into this area by examining specific development activities as pull request contributions in open-source software (OSS). This shows that the addition of features tends to increase TD Principal, refactoring activities have an almost consistent positive effect (reducing TD Principal), and bug fixing activities have an indecisive impact on TD Principal. Furthermore, we have focused on refactoring activities, since they seem to be the de facto way of reducing TD. From our results, it is evident that refactorings do not always positively affect the quality of the project, and that the list of refactoring opportunities can be way too long and costly to apply them all. Having this in mind, and by reusing a benchmark of real-world refactorings, we have proposed a metrics-based approach for guiding practitioners on how to prioritize refactoring opportunities. Following the second line of research, this PhD project has also provided an updated perspective on TD management for modern applications, like cloud and web services. First by acknowledging that in web services, developers organize the desired functional-ity into self-contained and independent services (API calls), it is only natural that the maintenance of them is independent. Thus, it makes sense to use as unit of analysis the service-per se, rather than the complete project or specific classes at a time. Alternative-ly, we proposed a novel methodology to measure the TD of each service considering the underlying code that supports the corresponding endpoint. The proposed methodology relies on the method call graph, initiated by the service end-point, and traverses all methods that provide the service functionality. Moreover, we proposed and present a cloud development methodology and framework, namely SmartCLIDE. This methodology and framework have in their core the reusability of services to be more efficient (in terms of time, quality, defects, and process). Finally, to provide a more concrete approach to software development, this PhD project expands to the design and architecture area of software engineering. Given the lack of metrics for the usage of libraries, we proposed 5 metrics that capture the extent to which each library is utilized in the code-base, providing information to the software engineers on the actual utility of the library in the final product.

Degree

thesis:*
Grantor dc:publisher
Πανεπιστήμιο Μακεδονίας
Year dc:date
2024

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Nikolaidis, Nikolaos
  • Νικολαΐδης, Νικόλαος

Subjects

dc:subject × 16

Rights

Language dc:language
eng

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/57696

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Nikolaidis, Nikolaos; Νικολαΐδης, Νικόλαος. Managing technical debt in modern software applications. Πανεπιστήμιο Μακεδονίας, 2024. http://hdl.handle.net/10442/hedi/57696