Procédés logiciels et Ingénierie des modèles

Journée de l'action IDM (GDR ASR & GPL)

23 juin 2010, Paris, France

UPMC, 4 place Jussieu, Tour 25-26 salle 105

Programme

9h30 : accueil

10h : L. Osterweil (University of Massachusetts)

Managing Resources to Support Process Execution

The performance of an executing process can be greatly affected by the resources that are available, and the way in which the resources are managed. In considering how to manage the resources needed by real-world processes, it is often necessary to address the need for very heterogeneous resource mixes, and highly dynamic utilization contexts. Significant technical problems arise in addressing these needs. Indeed the perspectives raised by these needs create some doubt about whether there is even a good understanding about what the inherent nature of a resource is. This talk summarizes some recent research aimed at developing a resource management system for use in supporting complex processes. This system is based upon novel conceptualizations of the inherent nature of resources, and the ways in which realistic systems and processes use them.
The talk begins with a very brief orientation to the Little-JIL process definition language, which is used as the basis for motivating a clearer understanding of the complexities of dealing with the resource needs of real-world processes. The talk then describes ROMEO, a prototype resource specification and management system used to support the management of resources used by processes defined in Little-JIL, and then presents some preliminary evaluations of ROMEO, and the resource management concepts upon which it is built. The talk then describes TWINS, a prototype resource scheduling system that builds upon the ROMEO architecture, and presents some preliminary evaluations of TWINS.

11h: J. Estublier (LIG)

Flexible process systems through model and metamodel composition

A process system must be adapted (or extended) to fit the needs of the business concepts, technology, and other specific functional and non functional characteristics of the domain in which it is be used. We have developed and experimented a system in which a process system is built by composition of a number of "domains". Each domain is defined by a metamodel (or domain model), an interpreter and a number of models. A process system is built around a process kernel, and composing its metamodel and model with an arbitrary number of additional domains. We will exemplify and demonstrate the resulting tool FOCAS.

12h-14h : déjeuner

14h: L. Clarke (University of Massachusetts)

Analyzing Process Definitions

Our work on systematic process improvement for human-intensive processes has demonstrated the need for extensive analysis capabilities combined with a process modeling language that provides a rich and well-defined set of semantic capabilities. This talk will describe several analysis techniques, including finite-state verification, fault-tree analysis, failure mode and effects analysis, and discrete-event simulation, as well as some process validation approaches. The techniques we applied found errors in the process models, helped improve understanding about the processes, and led to improved processes. This work has lead us to believe that modeling must be combined with rigorous analysis or the models will be too faulty to be the basis for important decision making.

14h30 : R. Lbath (IRIT)

IRIT-MACAO Works on Software Processes and MDE

In this presentation, I will briefly describe past, recent, and current research works conducted by our team MACAO (Modèles, Aspects, Composants pour des Architectures à Objets) at IRIT (Institut de Recherche en Informatique de Toulouse) in the domain of software processes and model-driven engineering.
First, I will present in a few words our past works achieved in ‘traditional’ (i.e. non-model-driven) software processes modeling. Then, I will present our recent works on process deviations management and demonstrate the approach we propose (logical-based detection of process deviations and rule-based decision-making mechanism). Finally, I will present our current works regarding model-driven process modeling and enactment by describing SPEM4MDE, a metamodel we developed for describing MDE processes structure and behavior.

15h: A. Canals et Philippe Desanois (CS Communication&Systèmes)

MDA-CS

Our presentation, is focused on our MDA process, called MDA-CS. First we talk about “How to be effective with the MDA approach”, choose a domain, a modelling tool …, configure the modelling tool, training teams, validate and adjust …, next we present an introduction to MDA-CS (based on UML-CS) through eleven transformations (from analysis to design, from design to tests, from design to presentation layer …), at last we provide a general overview, our feedback, the configuration efforts and our references.

15h30 : pause

15h45 : R. Bendraou (LIP6)

Early Deviation Detection in Modeling activities of MDE Processes

Software Process Models (SPM) are used to communicate around the process and to analyze it. They represent the entry point to PSEEs (Process-centered Software Engineering Environments) which use them to coordinate process agents in their tasks. When the process execution doesn't match the model, the common option provided by PSEEs is to ignore the model. If the actions of the agents are not tracked during deviations it is impossible to evaluate the effect of these deviations on the success or failure of the process. In this talk we present an approach for dealing with agent deviations during process execution. The originality of the approach comes from the fact that it is Process Modeling Language's (PML) independent and that it proposes early deviation detection.

16h15: discussion
17h15 : clôture