Back to search

Aristotle University Of Thessaloniki (AUTH)

Επαναχρησιμοποίηση λογισμικού βασισμένη στην ανάκτηση αρχιτεκτονικής πληροφορίας

Abstract

dc:description

This thesis focuses on methods that facilitate source code reuse. More specifically, it investigates methods to recover architectural and structural information in order to enable systems’ source code understanding by software engineers and facilitate source code reuse. Software reuse is the incorporation of libraries or software systems to other systems under development, different than the ones that were originally developed for. Software reuse has been established as a common software development practice, considering its inherent benefits that include reduced development effort and increased productivity. Free/Libre Open Source Software (FLOSS) repositories positively contribute to software reuse since the source code is available. Additionally, the enormous amount of FLOSS projects in repositories positively contributes to reuse practices. Although the benefits of software reuse are important, successful reuse is not ensured. A common phenomenon in several project categories, e.g. FLOSS, is the unavailability of technical or architectural documentation. Thus, their understanding lies on the comprehension of the source code. The information extraction and source code comprehension can become a difficult and time consuming process, especially for large and complex software systems.The aforementioned implications were the motivation for this thesis and the contributions are: Initially, the correlation between classes’ design attributes and their participation in architectural layers was investigated. In order to approximate the actual architectural layers, graph techniques were used and by extracting association rules, it was investigated if the design metrics are suitable to assign classes to the approximate architectural layers. Next, a reusability assessment metric was proposed that is based on classes’ participation in architectural layers. The goal of this metric is the fast and accurate calculation of the reusability of the systems’ elements, given that the elements that participate in lower architectural layers are more reusable compared to elements in higher layers. Next, an automated methodology was proposed in order to filter out noise during the analysis of software systems and to improve systems’ partitioning to subsystems. The methodology aims to identify those classes in a system, which import noise in software architecture recovery techniques. The methodology is based on the analysis of the graph that is derived by the software system, as well as on all the possible subgraphs that can be discovered in the system. The outcome was finding the classes/nodes in the system that if were removed, they would not negatively affect the existing software architecture recovery techniques. Next, a new method to recover the software architecture was proposed, which overcomes some of the problems that SAR techniques face. The method is based on combining different information sources and recent clustering algorithms. The proposed method improves the SAR results since it presents similar accuracy in the produced architectures, which is a problem of current SAR approaches.Next, a semi-automated process was proposed which aims to assist in understanding the structure of candidate reusable components from systems with complex structure due to cyclic dependencies. The process aims to decrease the candidate components’ sizes without affecting their provided functionality. In the context of this work, a method to identify the key nodes of the component is proposed, i.e. nodes that are responsible for highly increasing the component’s size. The method recommends and models the transformation of the key classes’ dependencies in order to simplify the dependencies between the component’s participants and enable its understanding. In a following step, the frequency and intensity of reuse practices was investigated in a large sample of FLOSS, as well as the modification of reuse decisions during projects’ evolution. The aim was to collect evidence of the actual effort that is required for software reuse during projects’ evolution. Next, a metrics suite was proposed in order to assess the architectural stability and evolution of software systems in the context of their reuse. The metrics quantified the architectural stability and evolution of software projects during their evolution. The metrics are considered more valuable for the reused library upgrades process since the provide evidence about possible implications during the upgraded version reuse.Overall, this thesis contributes to the comprehension of the design and architecture of software systems in order to facilitate software reuse.

Degree

thesis:*
Grantor dc:publisher
Aristotle University Of Thessaloniki (AUTH)
Year dc:date
2015

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Constantinou, Eleni
  • Κωνσταντίνου, Ελένη

Subjects

dc:subject × 12

Rights

Language dc:language
gre

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/36057

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

Constantinou, Eleni; Κωνσταντίνου, Ελένη. Επαναχρησιμοποίηση λογισμικού βασισμένη στην ανάκτηση αρχιτεκτονικής πληροφορίας. Aristotle University Of Thessaloniki (AUTH), 2015. http://hdl.handle.net/10442/hedi/36057