Walcélio L. Melo

Understanding and predicting the process of software maintenance releases

One of the major concerns of any maintenance organization is to understand and estimate the cost of maintenance releases of software systems. Planning the next release so as to maximize the increase in functionality and the improvement in quality are vital to successful maintenance management. The objective of the paper is to present the results of a case study in which an incremental approach was used to better understand the effort distribution of releases and build a predictive effort model for software maintenance releases. The study was conducted in the Flight Dynamics Division (FDD) of NASA Goddard Space Flight Center (GSFC). The paper presents three main results: (1) a predictive effort model developed for the FDDs software maintenance release process, (2) measurement-based lessons learned about the maintenance process in the FDD, (3) a set of lessons learned about the establishment of a measurement-based software maintenance improvement program. In addition, this study provides insights and guidelines for obtaining similar results in other maintenance organizations.

Characterizing and assessing a large-scale software maintenance organization

One important component of a software process is the organizational context in which the process is enacted. This component is ofien missing or incomplete in current process modeling approaches. One technique for modeling this perspective is the Actor-Dependency (AD) Model. This paper reports on a case study which used this approach to analyze and assess a large software maintenance organization. Our goal was to identify the approachs strengths and weaknesses while providing practical recommendations for improvement and research directions. The AD model was found to be very useful in capturing the important properties of the organizational context of the maintenance process, and aided in the understanding of the flaws found in this process. However, a number of opportunities for extending and improving the AD model were identified. Among others, there is a need to incorporate quantitative information to complement the qualitative model.

Software process model and work space control in the Adele system

It is argued that the most critical aspects for modeling and control, in a large software engineering environment, are inter/intrateam communication and synchronization. A solution based on a two-level approach is proposed. The Adele kernel supports multiple activities on shared objects, providing services like contextual behavior, active relationships, and general process support. The second level is the TEMPO formalism based on the role concept, which defines a software process step as a set of objects playing a role. Each objects characteristics and behavior depend on the role it plays in the software process it belongs to, and may be part of different simultaneous software processes. TEMPO clearly separates the description of the process from the description of the interaction and collaboration between the different processes.<<ETX>>

Supporting Software Development Processes in Adele 2

After years using of Adele, a configuration management system, it became apparent that it lacks activity- related concepts and mechanisms like work environment control, user coordination and synchronization, method and tool control, etc. It was also clear that considerable work is required to adapt a Software Engineering Environment (SEE) to user requirements. Using this experience, Adele 2 has been implemented to provide a general support for defining and managing dynamic aspects of a SEE and facilitate the building of new SEEs. This paper describes, using an example (work space control), the concepts and mechanisms involved. We show how close integration of an activity manager with a software engineering database fulfills the basic requirements and how a high level task manager coupled to a engineering database fulfills the basic requirements and how a high level task manager coupled to a configuration manager can be developed

Supporting software maintenance evolution processes in the Adele system

One of the major problems encountered when developing large systems is related to maintaining an operational and responsive software system, once it has been accepted and put into production. This problem is referred to as Software maintenance. Evolution is central to Software Maintenance, responsible for ensuring a longer working life. Many Software Engineering Environments (SEEs) have been constructed in order to support maintenance activities. In this paper, we will first present major development in SEEs to support Maintenance. Afterwards, we will describe the main ideas behind the design and implementation of the ADELE system, a third generation SEE. Following this, we will give an example to illustrate the way a process model can be described on top of ADELE using an example of evolution maintenance. This approach is heavily based on event-condition-action formalism. We will illustrate the way in which communication and coordination of the activities carried out by different users are automated.

Process centered SEE and Adele

The role and the evolution of software engineering environment (SEE) and computer-aided software engineering (CASE) environments in large software development and maintenance are discussed. The drawbacks of the current state of the art are emphasized. Improved structural and behavioral modeling are advocated, as is the introduction of team support to meet software engineering requirements. The different trends in SEE are addressed with emphasis on the evolution towards more powerful data modeling and the integration of process models.<<ETX>>

Cooperative work in large-scale software systems

Software maintenance is responsible for approximately 80% of the total cost of software. In order to reduce maintenance costs by improving both software productivity and quality, several software engineering environments have been built. In this paper, we present the main characteristics of the Adele system, showing how this system is able to take account of software maintenance problems. We concentrate on the ability of this system to accept the speci cation of project speci c policies and how these policies can be enforced automatically. We illustrate this with examples which deal with software maintenance. The second example is presented in the annex and shows an example of cooperative work in an aerospace company. We conclude this article by presenting the lessons learned using our approach and our projects for the future.

The need for a cooperative model: the Adele/Tempo experience

Our position is that the software process is fundamentally cooperative. We have been developing concepts and mechanism to tackle with cooperation in the Adele/Tempo project. Firstly, we integrated into the Tempo software process formalism a new concept: cooperation classes. In addition, we provided temporal, event-condition-action (TECA) rules in order to monitor and control cooperative activities. Based on TECA rules and cooperation classes, we can describe exchange information policies between software developers as well as rules about the organization, evolution and consistency of software projects. Secondly, we have been developing a transaction management system which is able to support the Tempo cooperative model. This system has been built up to deal with un-predicted events provoked by activities carried out in cooperative work environments.

The Object Role Software Process Model

In our previous papers[2, 1, 4, 3], we report our use of event-condition-action formalism for modeling software process, and the Adele trigger mechanism to control this execution. Although, this approach produced very good results, it has various drawbacks, including: (1) the formalism is low level; (2) trigger execution may be di cult to control; and (3) the software process description is distributed over di erent object and relation types. For human understanding, it would be preferable to have all the information relevant to a given process together in single unit. We are currently trying to solve the problems we found during our experiments using a high level process formalism. In the remainder of this position paper, we summarize the main design issues of this software process formalism.

Activity Coordination In Adele: A Software Production Kernel

A software production kernel aims to control and to help the software development and maintenance activities in an automated, integrated and consistent way. During the software life cycle, several activities are potentially automatic; a production environment must automate these tasks. By integrated we mean that an environment uses a central database managing the information and objects exchanged among the users. By consistent we mean that a software production environment must follows some predefined rules and policies. One of the main difficulties is that several activities are undertaken in parallel in different work spaces. The control and synchronization of work spaces is a key feature both for the users comfort and for consistency.