Fort de son expérience acquise au fil des années dans la mise en oeuvre de son système de traitement informatisé du contentieux et de la jurisprudence en droit du travail, l'unité de recherche en droit du travail du Centre de recherche de droit privé de l'Université Nancy 2 (CERIT-CRDP) entend, à travers ce projet, développer un programme analytique original sur le processus de l'argumentation judiciaire en droit du travail. Ce programme consiste à développer un dispositif de traitement informatique sur l'argumentation judiciaire sur la base d'un travail préalable d'identification, de recensement et de classification des arguments, raisons ou autres éléments invoqués oar les Chambres sociales des Cours d'appel dans la motivation de leurs décisions, pour justifier de la prémisse normative du raisonnement qu'elles mettent en oeuvre. L'objectif est de mettre en oeuvre un outil de recherche, qui permettra d'engager et poursuivre des travaux théoriques d'appronfondissement sur le thème central de l'argumentation judiciaire en droit du travail. Sa vocation essentielle est d'apporter une contribution française à la théorie de l'argumentation juridique ou judiciaire à partir de l'observation et la caractérisation du processus argumentatif effectivement à l'oeuvre dans le raisonnement des magistrats d'appel dans le domaine du droit du travail. Ce dispositif sera, par ailleurs, mis au service, en particulier, d'une amélioration de la qualité du raisonnement des juges en droit du travail.

The TACOS project proposes a component-based approach to specify systems with high safety requirements. The approach concerns the first steps of the system development life-cycle, from the requirements phase to the specification phase, and aims at using or adapting existing languages and tools. Land transportation systems will be taken as an application domain (LORIA has received many calls from the ”R´egion Lorraine” about the certification issue). These systems, which are both distributed and embedded, require to express functional properties as well as non functional-properties, in particular time constrained response and availability of required services. Our proposition is motivated by : 1. the increasing request by decision-makers for certification and confidence justification of the software parts in such systems, 2. the role of a rigorous approach to elicit requirements as a necessity to build systems which cover the needs of users and behave correctly with respect to the characteristics of their environment. The project aims at answering the following questions : - At the requirement level : how non-functional properties can be taken into account from the very beginning of the development life cycle ? Is it possible to combine the use of context diagrams introduced by Jackson and the KAOS requirements engineering method ? How to combine them with the use of UML diagrams ? - At the specification level : is it possible to use the B formal method with some extensions, like the addition of duration calculus, or in cooperation with other formalisms like CSP or timed automata ? Several SCIENTIFIC BOTTLENECKS have been identified in existing component-based approaches : 1. The identification of a relevant abstraction for the component expression. The specification of a component must include the complete information required to use it in a given context. A component can itself be defined as a combination of existing components. The information needed for the cooperation must be known. When is it possible to accept or reject the inclusion of a component in a system architecture ? This question is related to the interoperability problem : how can we ensure that the component will be correctly used by the client system, and that the required resources are available to the component ? 2. The need of high level languages to model components and their assembling, from the requirement level to the specification level. These languages should be rich enough to express assembly mechanisms between components, to take into account the verification of components compatibility and interoperability. Support tools are needed for the credibility of the approach. 3. Embedded systems. The chosen application domain involves embedded systems and softwares. Embedded systems entail time critical requirements for which very few studies have already been done. Some attempts to extend B with temporal logic allow us to address these real time specifications. 4. The traceability between the requirements model and the specification. Indeed, this notion of component applies to a software solution as well as to its requirements. This is still an investigation field to explore. The specification of a system consists in transforming the requirements model into a software architecture of the system. The problem is to find relevant rules that allow the translation of the elements of the requirements into elements of the system specification that form a set of well-fitted components. The TACOS project will focus on three main topics : 1. The expression of non-functional properties at the requirements level. 2. The specification of trustworthy component assembly : needs for the approach at the component specification level, assembly mechanisms and verification of the assembly. 3. The traceability between requirements and component specifications. The INNOVATIVE CHARACTER OF THE PROJECT lies on the use of components the specification level with a special emphasis on component assembly : description, verification and tools for proof. The challenge is to build trustworthy systems which satisfy the requirements and verifies important properties within the application domain of land transportation systems. Requirements have also to be formalised in order to build the right system. This approach gives assets for the certification problem : the problem of confidence is taken into account at the product level with the use of trustworthy components as well as at the process level with the trustworthy component assembly. THE SCIENTIFIC SPIN-OFF of the project can be summarised by : 1. The construction of trustworthy software systems from existing components. 2. An incremental approach to specify and verify component assembling. 3. The elicitation of non-functional requirements and their integration in the specification.