Identification and visualization of variability implementations in object-oriented variability-rich systems: a symmetry-based approach

Abstract

Most modern object-oriented software systems are variability-rich, despite that they may not be developed as product lines. Their variability is implemented by several traditional techniques in combination, such as inheritance, overloading, or design patterns. As domain features or variation points with variants are not a by-product of these techniques, variability in code assets of such systems is implicit, and hardly documented, hampering qualities such as understandability and maintainability. In this article, we present an approach for automatic identification and visualization of variability implementation places, that is, variation points with variants, in variability-rich systems. To uniformly identify them, we propose to rely on the existing symmetries in the different software constructs and patterns. We then propose to visualize them according to their density. By means of our realized toolchain implementing the approach, symfinder , we report on a threefold evaluation, (i) on the identified potential variability in sixteen large open-source systems and symfinder 's scalability, (ii) on measuring symfinder 's precision and robustness when mapping identified variability to domain features, and (iii) on its usage by a software architect. Results show that symfinder can indeed help in identifying and comprehending the variability of the targeted systems.

IDE-assisted visualization of indebted OO variability implementations

“…As the unchanged and changed parts can be abstracted in terms of variation points (vp-s) with variants [12], a symmetry-based identification approach has been previously proposed [15] and tooled with the the symfinder toolchain [8]. The approach also relies on the notion of density of symmetries in the code to reveal the variability organization (i.e., vp-s with an important number of variants at class or method level) [13] and display it in the shape of a graph. To fix the lack of precision of this first approach in symfinder [13], the usage relationships (type used as an attribute or method parameter) has been taken into account with the definition of a parameterized density measure that enables to automatically reveal fewer but more relevant hotspot zones concentrating variability implementations [9].…”

“…The approach also relies on the notion of density of symmetries in the code to reveal the variability organization (i.e., vp-s with an important number of variants at class or method level) [13] and display it in the shape of a graph. To fix the lack of precision of this first approach in symfinder [13], the usage relationships (type used as an attribute or method parameter) has been taken into account with the definition of a parameterized density measure that enables to automatically reveal fewer but more relevant hotspot zones concentrating variability implementations [9]. However, the resulting symfinder's graph visualization is then harder to interpret on large systems with two types of relationships being displayed at the same time.…”

“…Not following the Software Product Line (SPL) paradigm, many object-oriented (OO) systems implement their variability in a single codebase, using the traditional OO mechanisms (i.e., inheritance, overloading of methods and constructors, design patterns) [1]. The variability is then intertwined with the structure and implementation of the business code, and without traceability to the domain knowledge, the identification and understanding of these variability implementations are hindered [13,14]. Moreover, as the OO mechanisms may cause technical debt, variability debt [18] is likely to be introduced at the code level.…”

“…Some recent approaches such as symfinder and its extensions [8,9,13] detect potential variation points (vp-s) with variants in the targeted systems. Dense zones of detected symmetries have also be shown to represent interesting locations in terms of variability implementations.…”

Customizable visualization of quality metrics for object-oriented variability implementations

“…This is especially the case of object-oriented (OO) systems that often implement their variability in a single codebase, using the traditional OO mechanisms (i.e., inheritance, overloading of methods and constructors, design patterns) [12,27,82]. This absence of dedicated implementation mechanisms causes the variability to be intertwined with the implementation, hampering its identification, analysis, and understanding as there is no traceability with domain information [83,85]. Being completely dependent on mechanisms causing technical debt, such systems are prone to introduce variability debt [95] at the code level, calling for a solution to better identify and understand it.…”