Empirical Assessment of Two Approaches for Specifying Software Product Line Use Case Scenarios - SITE UNDER CONSTRUCTION

Abstract:

Modularity benefits, including the independent maintenance and comprehension of individual modules, have been widely

advocated. In this paper we use a multimethod technique, including designed experiments, to emprically evaluate such benefits in the

context of two approaches for specifying product line use case scenarios: PLUSS, which is based on annotations and leads to a nonmodular

design for features; and MSVCM, which relies on aspect-oriented constructs for improving feature modularity. After evaluating

these approaches through several case studies, we find out that MSVCM reduces feature scattering and improves scenario cohesion.

These results, in fact, suggest that evolving a product line specification using MSVCM requires only localized changes. On the other

hand, the results of six designed experiments reveal that MSVCM requires more time to bootstrap product line specifications and,

contrasting with the case studies results, reduces the time to evolve a product line specification only when the subjects have been well

trained and acquired enough experience with the task of evolving product line specifications.

Introduction

The evaluation of this study consisted in six case studies and three rounds of controlled experiments, each round containing two controlled experiments. This website makes available the content used to execute our evaluation studies.

Modularity Analysis' Case Studies

Here you can download the requirements' specifications used in our six case studies

Effort Analysis' Controlled Experiments

Our paper reports the execution of six controlled experiments organized in three rounds. Each round consisted of two controlled experiments using the same subjects, one experiment measures metrics to extract an SPL specification using MSVCM or PLUSS while the second experiment assess metrics related to evolving SPL specifications using MSVCM or PLUSS

Here you can download the material used during the three rounds of experiments.

If you have any questions please contact us using one of the following emails:

Rodrigo Bonifácio (rbonifacio@cic.unb.br)
Paulo Borba (phmb@cinNOSPAM.ufpe.br)
Cristiano Ferraz (cferraz@deNOSPAM.ufpe.br)
Paola Accioly (prga@cinNOSPAM.ufpe.br)

-- PaolaAccioly - 2013-08-09

MyForm
29 Jan.	EMN/ARMINES, ULANC, UMA, FCT/UNL
12 Feb.	HOLOS, EMN/ARMINES
26 Feb.

Attachments

Topic attachments
I	Attachment	Action	Size	Date	Who	Comment
zip	effort_analysis.zip	manage	9317.6 K	2013-09-17 - 19:31	UnknownUser
zip	modularity_analysis_specifications.zip	manage	1068.0 K	2013-08-27 - 14:41	UnknownUser

Topic revision: r4 - 2013-09-17 - PaolaAccioly