Andrea De Lucia

Recovering traceability links in software artifact management systems using information retrieval methods

The main drawback of existing software artifact management systems is the lack of automatic or semi-automatic traceability link generation and maintenance. We have improved an artifact management system with a traceability recovery tool based on Latent Semantic Indexing (LSI), an information retrieval technique. We have assessed LSI to identify strengths and limitations of using information retrieval techniques for traceability recovery and devised the need for an incremental approach. The method and the tool have been evaluated during the development of seventeen software projects involving about 150 students. We observed that although tools based on information retrieval provide a useful support for the identification of traceability links during software development, they are still far to support a complete semi-automatic recovery of all links. The results of our experience have also shown that such tools can help to identify quality problems in the textual description of traced artifacts.

Development and evaluation of a virtual campus on Second Life: The case of SecondDMI

Video games and new communication metaphors are quickly changing todays young people habits. Considering the actual e-learning scenarios, embedded in a fully technological enabled environment it is crucial to take advantage of this kind of capabilities to let learning process gain best results. This paper presents a virtual campus created using Second Life which provides four distinct types of virtual space: common student campus, collaborative zones, lecture rooms and recreational areas. Second Life environments and objects have been designed and programmed to support synchronous lectures and collaborative learning. The Second Life virtual world has also been equipped with supporting tools enabling students and teachers to navigate among multimedia contents. Second Life and an ad-hoc developed Moodle plug-in have been integrated to naturally enrich the environment with LMS services, exploiting this 3D world to increase the interaction and communication opportunities between teachers and students, and among students, principally favoring planned and unplanned social encounters. We have conducted an experiment involving university students aiming at evaluating Second Life synchronous distance lectures in the proposed learning environment. The evaluation has been conducted considering that, in a 3D multi-user virtual environment, learning is strongly related to the user perception of belonging to a learning community, as well as to the perception of awareness, presence and communication. The results of the evaluation are very positive.

Conditioned program slicing

Abstract Slicing is a technique to decompose programs based on the analysis of the control and data flow. In the original Weisers definition, a slice consists of any subset of program statements preserving the behaviour of the original program with respect to a program point and a subset of the program variables (slicing criterion), for any execution path. We present conditioned slicing, a general slicing model based on statement deletion. A conditioned slice consists of a subset of program statements which preserves the behaviour of the original program with respect to a slicing criterion for a given set of execution paths. The set of initial states of the program that characterise these paths is specified in the form of a first order logic formula on the input variables. We also show how slices deriving from other statement deletion based slicing models can be defined as conditioned slices. This is used to formally define a partial ordering relation between slicing models and to build a classification framework.

How to effectively use topic models for software engineering tasks? an approach based on genetic algorithms

Information Retrieval (IR) methods, and in particular topic models, have recently been used to support essential software engineering (SE) tasks, by enabling software textual retrieval and analysis. In all these approaches, topic models have been used on software artifacts in a similar manner as they were used on natural language documents (e.g., using the same settings and parameters) because the underlying assumption was that source code and natural language documents are similar. However, applying topic models on software data using the same settings as for natural language text did not always produce the expected results. Recent research investigated this assumption and showed that source code is much more repetitive and predictable as compared to the natural language text. Our paper builds on this new fundamental finding and proposes a novel solution to adapt, configure and effectively use a topic modeling technique, namely Latent Dirichlet Allocation (LDA), to achieve better (acceptable) performance across various SE tasks. Our paper introduces a novel solution called LDA-GA, which uses Genetic Algorithms (GA) to determine a near-optimal configuration for LDA in the context of three different SE tasks: (1) traceability link recovery, (2) feature location, and (3) software artifact labeling. The results of our empirical studies demonstrate that LDA-GA is able to identify robust LDA configurations, which lead to a higher accuracy on all the datasets for these SE tasks as compared to previously published results, heuristics, and the results of a combinatorial search.

On the Equivalence of Information Retrieval Methods for Automated Traceability Link Recovery

We present an empirical study to statistically analyze the equivalence of several traceability recovery methods based on Information Retrieval (IR) techniques. The analysis is based on Principal Component Analysis and on the analysis of the overlap of the set of candidate links provided by each method. The studied techniques are the Jensen-Shannon (JS) method, Vector Space Model (VSM), Latent Semantic Indexing (LSI), and Latent Dirichlet Allocation (LDA). The results show that while JS, VSM, and LSI are almost equivalent, LDA is able to capture a dimension unique to the set of techniques which we considered.

Automatic query reformulations for text retrieval in software engineering

There are more than twenty distinct software engineering tasks addressed with text retrieval (TR) techniques, such as, traceability link recovery, feature location, refactoring, reuse, etc. A common issue with all TR applications is that the results of the retrieval depend largely on the quality of the query. When a query performs poorly, it has to be reformulated and this is a difficult task for someone who had trouble writing a good query in the first place. We propose a recommender (called Refoqus) based on machine learning, which is trained with a sample of queries and relevant results. Then, for a given query, it automatically recommends a reformulation strategy that should improve its performance, based on the properties of the query. We evaluated Refoqus empirically against four baseline approaches that are used in natural language document retrieval. The data used for the evaluation corresponds to changes from five open source systems in Java and C++ and it is used in the context of TR-based concept location in source code. Refoqus outperformed the baselines and its recommendations lead to query performance improvement or preservation in 84% of the cases (in average).

Detecting bad smells in source code using change history information

Code smells represent symptoms of poor implementation choices. Previous studies found that these smells make source code more difficult to maintain, possibly also increasing its fault-proneness. There are several approaches that identify smells based on code analysis techniques. However, we observe that many code smells are intrinsically characterized by how code elements change over time. Thus, relying solely on structural information may not be sufficient to detect all the smells accurately. We propose an approach to detect five different code smells, namely Divergent Change, Shotgun Surgery, Parallel Inheritance, Blob, and Feature Envy, by exploiting change history information mined from versioning systems. We applied approach, coined as HIST (Historical Information for Smell deTection), to eight software projects written in Java, and wherever possible compared with existing state-of-the-art smell detectors based on source code analysis. The results indicate that HISTs precision ranges between 61% and 80%, and its recall ranges between 61% and 100%. More importantly, the results confirm that HIST is able to identify code smells that cannot be identified through approaches solely based on code analysis.

On integrating orthogonal information retrieval methods to improve traceability recovery

Different Information Retrieval (IR) methods have been proposed to recover traceability links among software artifacts. Until now there is no single method that sensibly outperforms the others, however, it has been empirically shown that some methods recover different, yet complementary traceability links. In this paper, we exploit this empirical finding and propose an integrated approach to combine orthogonal IR techniques, which have been statistically shown to produce dissimilar results. Our approach combines the following IR-based methods: Vector Space Model (VSM), probabilistic Jensen and Shannon (JS) model, and Relational Topic Modeling (RTM), which has not been used in the context of traceability link recovery before. The empirical case study conducted on six software systems indicates that the integrated method outperforms stand-alone IR methods as well as any other combination of non-orthogonal methods with a statistically significant margin.

Decomposing legacy programs: a first step towards migrating to client-server platforms

Abstract A solution to the problem of salvaging the past investments in centralised, mainframe-oriented software development, while keeping competitive in the dynamic business world, consists of migrating legacy systems towards more modern environments, in particular client–server platforms. However, a migration process entails costs and risks that depend on the characteristics of both the architecture of the source system and the target client–server platform. We propose an approach to program decomposition as a preliminary step for the migration of legacy systems. A program slicing algorithm is defined to identify the statements implementing the user interface component. An interactive re-engineering tool is also presented that supports the software engineer in the comprehension of the source code during the decomposition of a program. The focus of this paper is on the partition of a legacy system, while issues related to the re-engineering, encapsulation, and wrapping of the legacy components and to the definition of the middleware layer through which they communicate are not tackled.

Assessing IR-based traceability recovery tools through controlled experiments

We report the results of a controlled experiment and a replication performed with different subjects, in which we assessed the usefulness of an Information Retrieval-based traceability recovery tool during the traceability link identification process. The main result achieved in the two experiments is that the use of a traceability recovery tool significantly reduces the time spent by the software engineer with respect to manual tracing. Replication with different subjects allowed us to investigate if subjects’ experience and ability play any role in the traceability link identification process. In particular, we made some observations concerning the retrieval accuracy achieved by the software engineers with and without the tool support and with different levels of experience and ability.