Context: Through the years researchers have proposed several approaches to foster the growth and quality of experiments in Software Engineering. Among these approaches, there are some initiatives that rely on tool support for specifying controlled experiments. Goal: This paper reports results from a study, which aims to organize, analyze and outline the specifications of each initiative through a comparative analysis. Method: Specifications of each initiative have been compared through a comparative analysis, carried out according to eight criteria: (i) standard empirical concepts, (ii) goals and targets, (iii) involved variables, (iv) subject description, (v) design of experiment, (vi) tasks and activities, (vii) instruments and measurements, and (viii) the threats to research validity. Results: The results show that, among the tools currently existing and used in literature, the eSEE (Experimental Software Engineering Environment) is a complete model. However, it is also the most complex. In the other hand, the most flexible one is Experiment DSL. Conclusion: Based on our results, the currently existing solutions have strengths and weaknesses that should address efforts to make improvements in this area. In principal, our general suggestion is to place emphasis on methodological quality âĂŞ more than on method quantity.

Context: technology transfer and innovation dissemination are key success factors for an enterprise. The shift to a new software technology determines inevitable changes to ingrained and familiar processes and at the same time leads to changes in practices, requires training and commitment on behalf of technical staff and management. As so, it cannot leave out neither organizational nor technical factors. Objective: our conjecture is that the process of innovation dissemination is facilitated if the new technology is supported by empirical evidence. In this sense, Empirical Software Engineering (ESE) serves as support for transferring an innovation, either it being a process or product, within production processes. Method: this paper investigates the relation between empirical studies and technical/organizational factors in order to identify the most suitable empirical study for disseminating an innovation in an enterprise. The analysis has been carried out with respect to empirical studies carried out during a ten year time span within the Software Engineering Research LABoratory (SERLAB), at the University of Bari. Results: the results point out that a critical factor in designing empirical studies is the quality model, i.e. measurement program defined by researchers during the study and used to collect relevant information on the effectiveness and efficacy of the innovation being transferred, in order to gain commitment on behalf of stakeholders who will finally adopt the technology. Conclusion: the study outcomes provide an empirically founded guideline that can be used when choosing the most appropriate approach for addressing an innovation.

Abstract ISO has recently published Part 7 of the ISO/IEC 15504 standard, with the aim of determining the extent to which an organization consistently implements processes that contribute to achievement of its business goals. This new Part 7 of ISO/IEC 15504 has addressed issues related to the assessment of organizational maturity, as the CMMI‐DEV model also does. Recently, growing interest has been shown towards the need to harmonize different improvement models or standards, thereby presenting an integrated vision about them. All this being so and in an effort to offer information on how the maturity levels described in these two models are related, we have carried out the harmonization of these two models. It is based, firstly, on a mapping between processes of ISO 12207:08 and process areas of CMMI‐DEV, and in second place, on a matching between processes of ISO 12207:08 and ISO 15504‐5. For this work, we have taken into account the latest versions of the models, and defined a suitable process to carry out the mapping in a systematic way. We established differences and similarities between the maturity levels (and their processes) described in these models, our goal being to support organizations which are interested in tackling organizational maturity. Copyright © 2009 John Wiley & Sons, Ltd.

The knowledge deriving from the interactions between organizations offering a service/product and their users can be used to perceive new ideas for service/product improvement. Whenever customers communicate with an organization, unconsciously they give some improvement suggestions that can be extracted through storing, analysis and treatment of their requests, comments and observations. All this induces research studies towards an investigation of tacit suggestions extraction techniques, tools and processes. In this work a framework based on the use of a CRM and a tacit suggestions extraction process is proposed. The framework is applied to an E-Learning context in order to extract new suggestions for the E-Learning services improvement.

As a result of IT outsourcing and offshoring, IT professionals and educators are faced with the following question: What SE & KE skill sets will make a software engineer or a knowledge engineer immune to the impact of outsourcing and offshoring? This article summarizes the position papers from a panel held during the 2006 International Conference on Software Engineering and Knowledge Engineering from July 5 to 7 at the Hotel Sofitel, Redwood City in California, USA. Bringing software and knowledge engineers closer to the needs of their prospective customers and providing more value than simply pure software development and maintenance, is an open challenge at least for traditional computer science and software engineering curricula.

Objective: Evaluate the value of a product line in terms of maintainability, extensibility and configurability with refer to the interested stakeholders: customers, maintainers, producers. Rationale: There are values that customers constantly require in a modern software application. Some of these values are supported by product lines. Nevertheless, in the industrial and scientific communities the conjecture that customer values clash with those of producers/maintainers is diffused. Design of Study: we have designed and carried out a case study in an industrial context on an ongoing project to verify the validity of a product line in creating value for stakeholders. So data was collected as the project was being executed along a nine month period. Then, descriptive statistics and hypothesis testing were carried out. Results: experience acquired during the execution of an industrial project has allowed the authors to point out the differences between program families and software product lines. Also, the case study has shown how product lines contribute to stakeholder value proposition elicitation and reconciliation. Conclusions: This study has represented a first step towards analyzing the value that product lines represent for various stakeholders.

Effort estimation is a long faced problem, but, in spite of the amount of research spent in this field, it still remains an open issue in the software engineering community. In two previous works the authors proposed an approach named dynamic calibration for effort estimation of software projects. In this paper they present a tool named SPEED that implements the dynamic calibration approach

Dynamic calibration (DC), presented by the authors in previous works has proved to be a flexible approach for massive maintenance software project estimation, able to recalibrate an estimation model in use according to relevant process performance changes pointed out by the project manager. Nevertheless, it results quite subjective in its application and tightly based on manager experience. In this work the authors present an improvement of the approach based on the use of statistical process control (SPC) technique. SPC is a statistically based method able to quickly highlight shift in process performances. It is well known in manufacturing contexts and it has recently emerged in the software engineering community. In this work, authors have integrated SPC in DC as decision support tool for identifying when recalibration of the estimation model must be carried out. This extension makes DC less "person-based", more deterministic and transferable in its use than the previous version. The extended approach has been experimented on industrial data related to a renewal project and the results compared with both, a concurrent approach such as analogy based estimation and its previous version. The results are encouraging and stimulate further investigation.

Software reuse can become a key factor for improving and guaranteeing software quality, when adopted systematically all along the software process. The main characteristic of reuse-oriented processes is that they require a common repository for storing, searching and retrieving software modules. Moreover, reuse occurs systematically and is an integrated part of the process. Previous works of the same authors have empirically shown that the full reuse maintenance model (FRM) slows down quality degradation following to maintenance interventions on a software system. This work is a further step in the investigation towards demonstrating how reuse oriented development (ROD) impacts on software quality; how it favors FRM model; and finally, whether reuse-oriented development influences productivity, and as so, is more efficient. This has been done through a case study carried out on two ongoing industrial projects. Results are positive and support our research hypotheses.

Effort estimation is a long faced problem, but, in spite of the amount of research spent in this field, it still remains an open issue in the software engineering community. This is true especially in the case of renewal of legacy systems, where the current and well established approaches also fail. This paper presents an application of the method named dynamic calibration for effort estimation of renewal projects together with its experimental validation. The approach satisfies all the requirements of the estimation models. The results obtained by applying dynamic calibration are compared with those obtained with a competitor method: estimation by analogy. The results are shown to be promising although further experimentation on field is needed.

During its life, a legacy system is subjected to many maintenance activities, which cause degradation of the quality of the system: When this degradation exceeds a critical threshold, the legacy system needs to be reengineered. In order to preserve the asset represented by the legacy system, the familiarity with it gained by the system's maintainers and users, and the continuity of execution of current operations during the reengineering process, the system needs to be reengineered gradually. Moreover, each program needs to be reengineered within a short period of time. The paper proposes a reengineering process model, which is applied to an in-use legacy system to confirm that the process satisfies previous requirements and to measure its effectiveness. The reengineered system replaced the legacy one to the satisfaction of all the stakeholders; the reengineering process also had a satisfactory impact on the quality of the system. Finally, this paper contributes to validate the cause-effect relationship between the reengineering process and overcoming the aging symptoms of a software system.

Subcontracting is one of the challenges modern software engineering must face. The need for decision models to guide software engineers in identifying when it is advisable and how to divide a contract, which risks it involves and which actions should be adopted to control them, is continuously increasing. Many researchers have proposed and experimented with these decision models from various viewpoints. We feel the need for a method able to merge and formalize these experiences in order for them to be reused by the entire community. This work presents an approach for formalizing such experience. It also proposes a decision model for dividing a contract according to its characteristics and to the cooperating organizations. This decision model, together with others present in the literature, have been integrated and formalized in an experience package.

A transferable quality model must be general. This implies that only the high level characteristics can be transferred into different settings and that the refinement of the characteristics into metrics must be operated according to the context peculiarities. In spite of the amount of quality models and approaches to quality models definition presented in literature, there are no experiences reported showing the same quality model reused in different environments. This statement needs further investigation. The aim of this work is to present an experience on the field that involves two industrial projects in which the same quality model was used. The study confirms the need to modify the quality model, not only in the metrics but also in the measurement processes and in the interpretation of the resulting measures as side effects.

A large software project may be distributed over multiple sites when the organization believes that there are not enough people to staff a single collocated team. However, previous empirical research in the context of telecommunication organizations has shown that distance may increase cycle time and costs. We report on a large software massive maintenance project in the information systems domain, which in part has been carried out on a single site and in part across multiple sites of the same organization. We performed a comparative postmortem analysis of the two parts. Our results show that, with respect to cycle time and cost no significant differences exist among the distributed and collocated work. Indeed there is a significant difference in communication during the project. This implies that for massive maintenance activities the distribution over multiple sites can be really helpful.

This paper describes a process of gradual reengineering of the procedural components of a legacy system. The process is integrated and completed by the data reengineering process analyzed in a previous paper by the same authors. The proposed method enables the legacy system to be gradually emptied into the reengineered system, without needing to either duplicate the legacy system or freeze it. The process consists of evolving the legacy system components toward firstly a restored system and then toward the reengineered system. Meanwhile, the legacy system can coexist with both the restored and the reengineered parts. By the end of the process, a single system will be in existence: the reengineered one. The method has been applied to reengineer a real system and demonstrated its ability to: support gradual reengineering, maintain the system at work during the process, minimize the need to freeze maintenance requests, renew the operative environment of the reengineered system with respect to the legacy system and, finally, eliminate all the system's aging symptoms.

The paper presents a method for dynamic effort estimation, together with its supporting tool, and experimental validation on a renewal project of a very aged software system. Method characteristics such as dynamic tuning and fine granularity allow the tool to quickly react to process variations. The experimental validation shows how the combination of meaningful predictors and fine grain calibration is effective for understanding the enacted process and its implicit changes, or controlling the efficacy of explicit process changes. The study also confirms that the estimation model is process-dependent and then cannot be reused for other processes, albeit similar.

Abstract To improve a working process, one or more process components can be replaced by innovative ones. The literature provides a possible source of innovative process components. However, it is often difficult to evaluate whether a particular innovative process component satisfying an improvement goal can be introduced to replace a process component in the working software process to be improved. In previous works, the authors of this paper proposed a characterization and evaluation framework to address the above problem. To improve the usability of the framework, a scenario was defined for each question to help the user answer it methodically. The experiment presented in this paper aims to verify if a scenario‐based reading technique improves the usability of the framework. The results obtained show that the technique proposed makes assessment of which process components of a working process are candidates for making innovation more reliable. Moreover, the scenario‐based reading technique is user independent and only a little training may be sufficient to sensibly improve the effectiveness of use of the scenarios. The trustworthy results of this technique and its independence of the user's previous knowledge and ability ensure that it can be applied methodically. Copyright © 2001 John Wiley & Sons, Ltd.

Abstract This study generalizes some of the symptoms of ageing of a legacy system. Each symptom is specified by metrics and the results of the measurements made suggest what operations should be undertaken to renew the software. The study is based on retrospective analysis of data collected during the execution of a large renewal process of a very old legacy system. It therefore provides evidence of the expected efficacy of such renewal processes and can be used to decide how best to plan them and manage them in order to increase their efficacy. It can also be used to define the reengineering requirements to ensure long life to the system despite successive evolutions of the application and the operation domain. The metrics can provide a basis for monitoring a software system to ensure that its quality does not degrade to such an extent that the most costly and risky renewal processes then have to be performed to improve it. Finally, the paper points out the problems with renewal processes that still remain open. Copyright © 2001 John Wiley & Sons, Ltd.

This paper describes an explorative experimentation on the behaviour of developers using a predefined process versus that of those using self-defined ones. The results of the experimentation indicate that it is more advantageous, in terms of minimization of defects and execution effort, to require the developers to tailor the processes they will use rather than to ask them to use a predefined process. Moreover, assistance by the innovation promoter favours the transfer of experience from the organization to the individuals. Finally, quality goals, the availability of suitable tools and specialized training can motivate the use of innovative technologies the organization intends to adopt, even when the developers self-define the processes for their introduction.

The past decade (1987-97) has assisted a growing interest in reverse engineering and the consequent proliferation of papers and experience reports about this subject from both academic and industrial environments. Many types of reverse engineering process components (i.e. methods, algorithms, techniques, procedures, processes model, etc.) have been proposed. To be able to reuse effectively the published results, we need to characterize the process components, understand the similitude and differentiate them, and choose the ones to be used either alternatively or together in the same process. The Characterizing Reverse Engineering Process component methodology (CREP) is a systematic and defined approach to compare either processes or process components. It provides the definition of a characterizing evaluation schema specialized for the reverse engineering. The first phase of the CREP method formalises the process components, and constructs the process models; while, the second phase characterizes the process components, highlighting their most peculiar concepts and grouping the homogenous information

Proposes a metric for expressing the entropy of a software system and for assessing the quality of its organization from the perspective of impact analysis. The metric is called “structural information” and is based on a model dependency descriptor. The metric is characterized by its independence from the method of building the system and the architectural styles which represent it at the various levels of abstraction. It takes into account both the structure of the components at all levels of abstraction and the structure derived from the links between the different levels of abstraction. The use of this metric makes it possible to monitor changes in the quality levels of the system during its life-cycle and to determine which level of documentation is causing the degradation of the system. This enables adopting timely measures to counter any drop in the quality of a system due to poorly-performed maintenance. In addition, the metric can serve to check that perfective maintenance objectives are attained

This study is based on the Structured Process Model formalism previously proposed by the same authors, and describes some techniques for scheduling an execution plan which conforms to the chosen process but adapts to all the potential changes in the value of specific variables of the project. The executive plan is generated by means of Incremental Refinements. Each increment is the consequence of one or more decisions taken by the manager. All the contingencies which can occur during the execution of the project are classified in a finite number of types of events which form the basis for the definition of all the transition laws between the possible states of the process. Naturally, the manager's decisions, and therefore the composition of the increment, take into account the changes in the value of the variables of the process and of the process model itself. In addition, the study proposes the handling of fuzzy activities, by means of which the problems of recycles and reprocessing are solved. These activities also provide a better management of the execution times of the project.

An alternative approach to developing reusable components from scratch is to recover them from existing systems. In this paper, we apply program slicing, a program decomposition method, to the problem of extracting reusable functions from ill-structured programs. As with conventional slicing first described by Weiser, a slice is obtained by iteratively solving data flow equations based on a program flow graph. We extend the definition of program slice to a transform slice, one that includes statements which contribute directly or indirectly to transform a set of input variables into a set of output variables. Unlike conventional program slicing, these statements do not include either the statements necessary to get input data or the statements which test the binding conditions of the function. Transform slicing presupposes the knowledge that a function is performed in the code and its partial specification, only in terms of input and output data. Using domain knowledge we discuss how to formulate expectations of the functions implemented in the code. In addition to the input/output parameters of the function, the slicing criterion depends on an initial statement, which is difficult to obtain for large programs. Using the notions of decomposition slice and concept validation we show how to produce a set of candidate functions, which are independent of line numbers but must be evaluated with respect to the expected behavior. Although human interaction is required, the limited size of candidate functions makes this task easier than looking for the last function instruction in the original source code.

A typical approach to software maintenance is analyzing just the source code, applying some patches, releasing the new version, and then updating the documentation. This quick-fix approach usually leads to documentation not aligned with the current system and degrades the original system structure, thus rendering the evolution of the system costly and error-prone. Although there are alternative maintenance models which avoid these problems, by analyzing and updating the system documentation first, the quick-fix approach continues to be popular because of the time pressure for new releases and the resistance to change of maintenance programmers. In this paper, we propose an iterative reengineering model which can be run each time the maintainability and reliability of a software system degrade under a tolerance level. The reengineering process, applied after a number of modifications, can result in renovation of the current system or simply in realignment of the documentation. In this context, reengineering is no longer a one-shot process but becomes an ordinary process which runs concurrently with the quick-fix maintenance process. The results obtained with an industrial case study are presented and the lessons learned are discussed.

Software evolution has no common paradigm which practitioners can adhere to. On the contrary, there is a wide range of models, methods, techniques, and tools which are selected according to the specific task, the application domain, the professional experience, and the organizational culture. We argue that different approaches and technologies may be combined into a unique platform to satisfy the needs of software systems which evolve over long periods of time. This paper presents the Integrated Environment for Software Evolution Management (IESEM) which includes software repositories, reverse engineering tools, rationale capture tools, software measurement tools, and a user-friendly interface. It can manage heterogeneous systems characterized by various design methods and programming languages. IESEM is based on a central repository which stores software engineering artifacts, program code, design, and implementation decisions in the form of a traceability graph. The repository stores also software measures computed both from programs and external CASE repositories. Measures are used to control software degradation during its evolution and to support decisions based on quality factors. The key concepts of IESEM, its design, and implementation are presented. The use of IESEM during development and maintenance is discussed. A case study shows IESEM's effectiveness in performing maintenance tasks.

Abstract The destiny of legacy software is a relevant economic problem. Many companies are experiencing the friction that legacy systems oppose to the change and growth of their business. Nevertheless, these systems cannot be simply discarded and replaced with new ones developed according to the new organization requirements. Existing systems record a great deal of knowledge and expertise used to set up solutions to real problems in different application domains and it is imperative that this knowledge must not be lost, partly because it may not be recorded anywhere else than in the code. This paper discusses the problems related to the modularization of legacy systems, with the emphasis being placed on the automation of code scavenging. It is argued that a balance must be struck between the modularization of legacy code, and the production of reusable, easy‐to‐evolve modules. The paper provides a framework for examining system modularization methods and identifies a set of attributes that define their quality. An overview of modularization techniques is presented, and an evaluation is made to assess their strength and identify and understand their major limitations. Finally, an attempt is made to identify techniques and tools that are nature enough to be transferred from research laboratories to industry.

Abstract This paper presents our approach to design recording which aims to facilitate the impact analysis of changes in data, functions, or the external environment. A whole software system is represented as a web which integrates the different work products of the software life cycle and their mutual relationships. A traceability relationship associates the objects with each other so that impact analysis can be performed. Internal traceability is provided by semantic links between software objects representing the work products of a development phase, while external traceability is assured by the cognitive links between software objects from different phases. System understanding is supported by the decisions which are involved in the transformation process. The history of these decisions is retained over time so that previous decisions can be examined for maintenance and reuse activities. The approach has been implemented through a Traceability Support System, a maintenance tool which combines the characteristics of program abstractors, project databases and design rationale capture tools. The approach and the tool also both support traceability in heterogeneous systems, which have subsystems implemented on different platforms. Finally, analysis is made of the results of an empirical investigation carried out to assess the approach.

An experimental study is presented in which participants perform impact analysis on alternate forms of design record information. The primary objective of the research is to assess the software maintainer's performance with respect to various approaches of design recording. One of the approaches is the model dependency descriptor, which includes decision capturing and explicit traceability links between software objects and decisions. Results indicate that design recording approaches slightly differ in work completeness and time-to-finish, but the model dependency descriptor leads to the most accurate impact analysis. These results suggest that design records have the potential to be effective for software maintenance, but training and process discipline is needed to make design recording worthwhile.>

Some features of software create considerable problems as regards the management of its construction processes, which therefore requires suitable models. In the studies published up to now on this subject, various problems have remained unsolved, particularly the readability of the numerous independent components for each activity in the process, the help required by the manager for handling projects and the need to adapt models to risks and the software development environment. The model proposed in this paper offers a solution to these problems. The innovative aspects of the approach described are the location of all pertinent information in the products of the model in order to improve management, and a formal definition which makes it possible to control the internal coherence of the plan automatically using a flexible notation which is compatible with the knowledge of all participants in the production process.

The identification and extraction of two main kinds of components, environment-dependent operations and domain-dependent functionalities, are proposed. A reference information model drives the process by giving expectations of components and their interface data. Two modified definitions of Weiser's slicing are applied to this function recovery problem: direct slice and transform slice. Direct slice is an executable subset of the original program containing all the statements which directly contribute either to the writing on an external sink or to the reading from an external source. Transform slice is also an executable subset including all the instructions which directly or indirectly contribute to transform an external input into an external output.>

The authors propose the use of interactive animation techniques as a support to reverse engineering processes oriented to the synthesis of semantic abstractions. Starting from data flow diagrams, a formal model, called dynamic data flow diagrams (DDFDs), has been defined, which can be used for the production of executable models of a software system. A strategy for the DDFD interactive animation is also put forward. Finally, the authors describe a prototype tool for (i) the production of the DDFD which models an ADA system starting from the analysis of the code and (ii) the interactive animation of such a model according to the proposed strategy.>

The authors model the different points of view on a software structure as distinct conceptual models. The mapping among models of the same software structure is made through the assumption of design decisions. This mapping is represented as a graph, the model dependency descriptor, which connects components of source models with those in the target models. A traceability relation is set between components by producing the transitive closure of the model dependency descriptor. The traceability relation which is proposed links objects and also tracks decisions having a role in the transformation. A traceability support system incorporates a number of design databases and implements the model dependency descriptor. Its use enables the management of systems which were developed by using different design methods and different programming environments, and the capturing of the assumptions set during both development and maintenance.>

The spur of innovation coming from new technologies and methodologies often leads to poorly integrated systems which evolve with no disciplined management model. The authors suggest a set of conceptual models for representing a system structure, differing as regards abstraction degree (essential vs. language-oriented), decomposition method (functional vs. object-oriented), and target languages. A set of software metrics is applied to system components, to single out the critical areas which require justification. Comparing measures between different versions, variants and degrees of abstraction, enables one to monitor the growth of entropy during software evolution.>

An environment is described which supports the reengineering cycle at the design level and provides a homogeneous software configuration for the maintenance process. The environment is based on an integration tool which takes the output from a static analyzer, used for reverse engineering and populates the repository of a computer-aided-software-engineering (CASE) tool, usually used in forward engineering. The integration allows enhancement of the knowledge of the program, the reliability of modifications, the reusability and the change management. A case study is shown in which the design of a sample program is recaptured from the source code.>