Automated error analysis for the agilization of feature modeling

doi:10.1016/j.jss.2007.10.030

Journal of Systems and Software

Volume 81, Issue 6, June 2008, Pages 883-896

https://doi.org/10.1016/j.jss.2007.10.030 Get rights and content

Abstract

Software Product Lines (SPL) and agile methods share the common goal of rapidly developing high-quality software. Although they follow different approaches to achieve it, some synergies can be found between them by (i) applying agile techniques to SPL activities so SPL development becomes more agile; and (ii) tailoring agile methodologies to support the development of SPL. Both options require an intensive use of feature models, which are usually strongly affected by changes on requirements. Changing large-scale feature models as a consequence of changes on requirements is a well-known error-prone activity. Since one of the objectives of agile methods is a rapid response to changes in requirements, it is essential an automated error analysis support in order to make SPL development more agile and to produce error-free feature models.

As a contribution to find the intended synergies, this article sets the basis to provide an automated support to feature model error analysis by means of a framework which is organized in three levels: a feature model level, where the problem of error treatment is described; a diagnosis level, where an abstract solution that relies on Reiter’s theory of diagnosis is proposed; and an implementation level, where the abstract solution is implemented by using Constraint Satisfaction Problems (CSP).

To show an application of our proposal, a real case study is presented where the Feature-Driven Development (FDD) methodology is adapted to develop an SPL. Current proposals on error analysis are also studied and a comparison among them and our proposal is provided. Lastly, the support of new kinds of errors and different implementation levels for the proposed framework are proposed as the focus of our future work.

Section snippets

Introduction and motivation

The so-called agile methods have arisen to face up to the problems that traditional, heavyweight software development methodologies have not satisfactorily solved yet. Agile methods pursue some main goals which are described in the Agile Manifesto (Fowler and Highsmith, 2001), where a number of key changes to traditional software development are proposed. For example: focusing the efforts during development in the interaction with customers through working software; collaborating with customers

Feature models

As briefly commented in Section 1, feature models are a widely used notation to describe the set of products in a software product line in terms of features. In feature models, features are hierarchically linked in a tree-like structure and are optionally connected by cross-tree constraints. An example on how feature models are depicted is shown in Fig. 2, where the feature model describes a Home Integration System (HIS) product line.

Although there are many proposals on the type of

Feature model level: dealing with errors in feature models

A feature model is composed by features and relationships among them. A feature model describes the products in a SPL, considering products as sets of selected features. Relationships are added to reduce the set of products until the SPL is properly described.

Sometimes, introducing new relationships in a feature model may accidentally remove some products so the feature model does not describe the real SPL. On the other hand, the feature model may not be correctly constrained so some products

Diagnosis level: diagnosing and explaining errors

The goal of the diagnosis level, as depicted in Fig. 1, is transforming a feature model into a diagnosis model in order to detect errors and provide their corresponding explanations. This transformation can be described using a circuit-like representation of feature models, where each relationship corresponds to a component. Every component or relationship has one binary input per feature and one binary output (see Fig. 7). Each input represents the presence (1) or absence (0) of a feature,

Implementation level: modeling diagnosis problem as a CSP

One of the main advantages of defining error detection and explanation in terms of theory of diagnosis is having the problem described in an implementation-independent way. In this work, we propose an implementation based on constraint programming, however, any other implementation could be proposed relying on the previous diagnosis level.

Our proposal is inspired by two main sources: on the one hand, Benavides et al. (2005) proposed a direct mapping from a feature model onto a CSP to extract

Applying our proposal to a real case

We have applied the implementation level described in this paper during a Software Product Line development project. The project intends to build a set of Enterprise Resource Planning (ERP) products in the context of SAUCE, an environmental resources management SPL. SAUCE comprises a set of products to store and exploit the existing information about flora and fauna in different rivers. The aim of the SPL is to produce customized software that helps to the management and conservation of these

Related work

Although the automated error analysis in feature models was already identified as a fundamental task in the original FODA report by Kang et al. (1990), few authors have dealt with it. As a matter of fact, there has not been a seminal approach to automatically analyze errors in feature models as far as we know.

Our interest in automating error detection and explanation arose from the work of von der Massen and Lichter (2004), where the authors proposed a categorization of what they call

Conclusions and future work

We have discussed our vision on how SPL and agile methods can come together, either by applying agile principles to SPL methodologies or by tailoring an existing agile methodology to support SPL development. Independently from the chosen alternative, supporting automatic error detection and explanation is an important contribution that can be a first step in bringing agile principles and SPL together. As feature modeling is an error-prone task, an activity that checks the feature model is

Acknowledgements

We would like to thank Richard Page, Patrick Heymans and Pierre-Yves Schobbens for their useful comments and Manuel Nieto for his effort on applying our ideas in an industrial case study.

Pablo Trinidad is a lecturer at the Department of Computer Languages and Systems of the University of Seville since 2004, where he received his MSc in Computer Science in 2002 (with honours). He worked on several consultant companies before joining the University of Seville. His main research interest is software product lines, having published several articles in international conferences.

References (27)

R. Reiter
A theory of diagnosis from first principles
Artif. Intell.
(1987)
P. Schobbens et al.
Generic semantics of feature diagrams
Comput. Network
(2007)
Batory, D., 2005. Feature models, grammars, and propositional formulas. In: Software Product Lines Conference, LNCS...
D. Batory et al.
Scaling step-wise refinement
IEEE Trans. Software Eng.
(2004)
D. Batory et al.
Automated analysis of feature models: challenges ahead
Commun. ACM
(2006)
Benavides, D., Ruiz-Cortés, A., Trinidad, P., 2005. Automated reasoning on feature models. In: LNCS, Advanced...
Benavides, D., Segura, S., Trinidad, P., Ruiz-CortTs, A., 2007. FAMA: tooling a framework for the automated analysis of...
Czarnecki, K., Kim, P., 2005. Cardinality-based feature modeling and constraints: a progress report. In: Proceedings of...
K. Czarnecki et al.
Model-driven software product lines
de Kleer, J., Mackworth, A.K., Reiter, R., 1990. Characterizing diagnoses. In: AAAI, pp....

Fattah, Y.E., Dechter, R., 1995. Diagnosing tree-decomposable circuits. In: IJCAI, pp....

Fowler, M., Highsmith, J., 2001. The agile manifesto. In: Software Development, Issue on Agile...

Haarslev, V., Moller, R., 2001. Description of the RACER system and its applications. In: Description Logics. URL:...

Cited by (101)

Evolutionary Computing to solve product inconsistencies in Software Product Lines
2022, Science of Computer Programming
In Software Product Lines (SPLs), multiple design teams work collectively to configure products. Often, having multiple sub-designs leads to inconsistencies which prevent and delay the configuration of the final product at the end. Solutions are available to identify and resolve these inconsistencies for limited scale SPLs. For large scale and industrial SPLs, resolution of inconsistencies is still a potential research topic. In this paper, we propose, implement and validate an application of evolutionary computation, particularly genetic algorithms, to resolve inconsistencies in large scale SPL products. We label our approach ga-SPLIS (genetic algorithm-Software Product LIne Solution). Using a concrete experimental methodology with different feature set sizes and multiple standard performance metrics, we generate results which validate the applicability of ga-SPLIS for the SPL community.
DIRECTDEBUG: A software package for the automated testing and debugging of feature models[Formula presented]
2021, Software Impacts
Complex and large-scale feature models can become faulty, i.e., do not represent the expected variability properties of the underlying software artifact. In this paper, we propose the DirectDebug algorithm that supports the automated testing and debugging of variability models. Our approach assists software engineers in identifying an adaptation hint (diagnosis) that makes all test cases consistent with the knowledge base. We also develop the software package so-called d2bug_eval to evaluate the DirectDebug’s performance. The software package can be re-produced thoroughly to evaluate consistency-based algorithms.
Classifying and resolving software product line redundancies using an ontological first-order logic rule based method
2021, Expert Systems with Applications
Software product line engineering improves software quality and diminishes development cost and time by efficiently developing software products. Its success lies in identifying the commonalities and variabilities of a set of software products which are generally modeled using feature models. The success of software product lines heavily relies upon the quality of feature models to derive high quality products. However, there are various defects that reduce profits of software product line. One of such defect is redundancy. While the majority of research work focuses on the identification of redundancies, their causes and corrections have been poorly explored. Causes and corrections must be as accurate and comprehensible as possible in order to support the developer in resolving the cause of a redundancy. This research work classified redundancies in the form of a typology. An ontological first-order logic rule based method is proposed to deal with redundancies. A two-step process is presented for mapping model to ontology based on predicate logic. First-order logic based rules are developed and applied to the generated ontology for identifying redundancies, their causes and corrections to resolve redundancies. The proposed method is illustrated using a case study from software product lines online tools repository. The results of experiments performed on 35 models with varied sizes of real world models as well as automatically-generated models from the Software Product Line Online Tools repository and models created via FeatureIDE tool conclude that the method is accurate, efficient and scalable with FM up to 30,000 features. Thus, enables deriving redundancy free end products from the product line and ultimately, improves its quality.
FCA for software product line representation: Mixing configuration and feature relationships in a unique canonical representation
2020, Discrete Applied Mathematics
Software Product Line Engineering (SPLE) is a set of methods to help build a collection of software systems which are similar enough to enable appropriate artefact reuse. An important task consists in documenting in variability models the common and variable features which may compose the similar software systems along with compatibility constraints between these features. Several models and formalisms have been proposed to model variability: each one of them has specific properties making it pertinent to support certain management operations, which are of primary importance in SPLE. Switching from one kind of variability model to another is thus important to benefit from a wide range of operations and efficiently manage a software product line. In this paper, we review the various approaches proposed to manage and organise features and product configurations (a product configuration being a chosen subset of features). We discuss the originality of concept lattices, canonical structures presenting a dual view on features and configurations, and the advantages to use these structures as variability representations. Switching from existing variability models to an equivalent concept lattice raises scaling issues related to the size of the needed input dataset and thus hinders their exploitation. We propose an alternative relying on implicative systems and define a transformation method which does not suffer from scaling issues.
Change impact analysis for evolving configuration decisions in product line use case models
2018, Journal of Systems and Software
Product Line Engineering is becoming a key practice in many software development environments where complex systems are developed for multiple customers with varying needs. In many business contexts, use cases are the main artifacts for communicating requirements among stakeholders. In such contexts, Product Line (PL) use cases capture variable and common requirements while use case-driven configuration generates Product Specific (PS) use cases for each new customer in a product family. In this paper, we propose, apply, and assess a change impact analysis approach for evolving configuration decisions in PL use case models. Our approach includes: (1) automated support to identify the impact of decision changes on prior and subsequent decisions in PL use case diagrams and (2) automated incremental regeneration of PS use case models from PL use case models and evolving configuration decisions. Our tool support is integrated with IBM Doors. Our approach has been evaluated in an industrial case study, which provides evidence that it is practical and beneficial to analyze the impact of decision changes and to incrementally regenerate PS use case models in industrial settings.
An approach for optimized feature selection in large-scale software product lines
2018, Journal of Systems and Software
Context: Feature selection in product line engineering is an essential step for individual product customization, in which the multiple objectives, that are often competing and conflicting, have to be taken into consideration. These objectives always need to be balanced during selection, leading to a process of multi-objective optimization. What’s more, the massive complex dependency and constraint relationships between features present another huge challenge for optimization.
Objective: In this work, we propose a multi-objective optimization algorithm, IVEA-II, to automatically search through configurations to obtain an optimal balance between various objectives. Additionally, all the relationships between features must be conformed to by the optimal feature solutions.
Method: Firstly, a two-dimensional fitness function in our previous work is reserved. Secondly, to prevent the negative impact of this 2D fitness on the diversity of final Pareto Fronts, the crowding distance is introduced into each fitness-based selection. Lastly, a new mutation operator is designed to improve the scalability of IVEA-II.
Results: A series of experiments were conducted to verify the effectiveness of IVEA-II on five large-scale feature models with five optimization goals.
Conclusion: Experiments showed that IVEA-II can generate more valid solutions over a set period of time, with optimal solutions also having better diversity and convergence.

View all citing articles on Scopus

David Benavides is a senior lecturer at the Department of Computer Languages and Systems of the University of Seville, where he received his MS degree in Computer Science in 2001 and his PhD in 2006. His research interests are software product lines and software engineering for web applications. He has published several articles and papers about software product lines and feature models analysis.

Amador Durán is an associate professor at the Department of Computer Languages and Systems of the University of Seville, where he received his MS degree in Computer Science in 1993 and his PhD in 2000. Currently, he is the vice dean of International Relationships of his faculty. His research interests are requirements engineering, business modelling, software product lines and software engineering for web applications. He has published several articles in international journals, several papers in conferences like the IEEE Requirements Engineering, and edited a book about requirements engineering.

Antonio Ruiz-Cortés is an associate professor of software engineering and member of the Department of Computer Languages and Systems from the University of Sevilla since 1998. He received the B.Sc (with honours), the M.Sc and the Ph.D. (with honours) in Computer Science from the same University. His current research interests include software engineering (Requirements Engineering, Process Modelling, Software Factories, Product Lines, Model Driven Development, Autonomic Computing) and service oriented computing (Service Oriented Architecture, Service Level Agreements, web services composition). Before joining the University, he worked as a software architect, requirements engineer and finally as a project manager for a leading environmental software firm. Currently, he is a member of the European Commission Board of Experts in Software Engineering and head of the Applied Software Engineering Group (http://www.isa.us.es).

Miguel Toro is a professor of computer science who received his PhD degree in electrical engineering in 1988. His current research activities cover a broad range of areas, including software engineering, formal methods, data mining, and qualitative reasoning. He has also led several national research projects on these topics.

^☆: This work has been partially supported by the European Commission (FEDER) and Spanish Government under CICYT project Web-Factories (TIN2006-00472).

View full text

Automated error analysis for the agilization of feature modeling☆

Abstract

Section snippets

Introduction and motivation

Feature models

Feature model level: dealing with errors in feature models

Diagnosis level: diagnosing and explaining errors

Implementation level: modeling diagnosis problem as a CSP

Applying our proposal to a real case

Related work

Conclusions and future work

Acknowledgements

Artif. Intell.

Comput. Network

Scaling step-wise refinement

IEEE Trans. Software Eng.

Automated analysis of feature models: challenges ahead

Commun. ACM

Model-driven software product lines