A Framework Based on Implementation Relations for Implementing LOTOS SpecificationsLeduc, Guy ![]() in Computer Networks & ISDN Systems (1992), 25(1), 23-41 A framework is developed for studying the implementation process, as a stepwise process in which an abstract specification is successively transformed to reach a final compilable specification adapted to ... [more ▼] A framework is developed for studying the implementation process, as a stepwise process in which an abstract specification is successively transformed to reach a final compilable specification adapted to the computer environment. In this context, an implementation relation is referred to as the relation which should link any ``valid'' implementation to its abstract formal specification. In other words, the implementation relation is intended to express formally the notion of validity. Our framework allows the exact characterization of the transformations which may take place at each step for a given implementation relation. This framework is essential for dealing with non-transitive implementation relations. In the second part of the paper, these results are exemplified in LOTOS on some existing relations, and an apparent paradox is presented. Some new results about these relations are also derived. [less ▲] Detailed reference viewed: 9 (0 ULg) OSI 95 Enhanced Transport Facilities and Functions; ; Leduc, Guy et alReport (1992) Detailed reference viewed: 10 (2 ULg) An Upward Compatible Timed Extension to LOTOSLeduc, Guy ![]() in Parther, Ken; Rose, Gordon (Eds.) Formal Description Techniques IV (1992) We propose a timed extension of LOTOS, denoted TLOTOS, which is upward compatible with the standard LOTOS. It means first that a timed behaviour expression which does not use any language extension will ... [more ▼] We propose a timed extension of LOTOS, denoted TLOTOS, which is upward compatible with the standard LOTOS. It means first that a timed behaviour expression which does not use any language extension will have the same semantics as in standard LOTOS. In addition, we have pushed this upward compatibility one step beyond by requiring and obtaining that all the familiar equivalence laws of standard LOTOS be preserved e.g. B [] B approximately B, B [] stop approximately B. This is needed in order to keep the intuition of a standard LOTOS user unchanged. TLOTOS has been mainly inspired by other approaches such as Moller Tofts's TCCS (Temporal CCS), Hennessy Regan's TPL (Temporal Process Language) and Nicollin Sifakis's ATP (Algebra of Timed Processes). Our model is not strictly asynchronous (like standard LOTOS, CCS, CSP, ACP) nor strictly synchronous (like SCCS, CIRCAL, Meije). For compatibility and simplicity, we decided to keep the model asynchronous for a large part and, for dealing with time, to introduce a `'synchronous part'' by way of a new basic `'synchronous'' or timed action in the asynchronous semantic model. The design of our TLOTOS is presented together with other possible design alternatives which are all rejected according to our compatibility or expressive power requirements. The solution that we obtain is simple and satifactory, and its operational semantics is presented in depth. Two examples are provided to illustrate the use of TLOTOS. [less ▲] Detailed reference viewed: 29 (17 ULg) A LOTOS Data Facility Compiler (DAFY); Leduc, Guy ![]() in Parther, Ken; Rose, Gordon (Eds.) Formal Description Techniques IV (1991, November) If we take a look at existing LOTOS specifications, we notice that the description of the needed data types is very often huge. This causes the lack of concision of most descriptions of complex systems ... [more ▼] If we take a look at existing LOTOS specifications, we notice that the description of the needed data types is very often huge. This causes the lack of concision of most descriptions of complex systems. We propose to tackle this problem in two steps. First, we define extensions to the LOTOS language allowing short definitions of most of the data types used in practical LOTOS specifications. Second, we propose a tool called `'DAFY'' (Data Facility Compiler) which is able to translate these extensions into standard LOTOS. [less ▲] Detailed reference viewed: 12 (0 ULg) Equivalence associée à la relation de conformité "conf" et simplification du testeur canonique en LOTOSLeduc, Guy ![]() in Rafiq, Omar (Ed.) CFIP'91 Ingénierie des Protocoles (1991, September) Detailed reference viewed: 12 (0 ULg) Conformance relation, associated equivalence, and minimum canonical tester in LOTOSLeduc, Guy ![]() in Jonsson, B.; Parrow, J.; Pehrson, B. (Eds.) Protocol Specification, Testing, and Verification, XI (1991, June) We first study the conf relation proposed by E. Brinksma and G. Scollo to formalise testing conformance. It is well-known from their work that, in order to test whether an implementation of a ... [more ▼] We first study the conf relation proposed by E. Brinksma and G. Scollo to formalise testing conformance. It is well-known from their work that, in order to test whether an implementation of a specification S (i.e. I conf S), it suffices to build, from S, a canonical tester T(S) such that, when T(S) is synchronised with an implementation I, it always reaches a correct final state if, and only if, I conf S. For instance, if I is not a valid implementation of S, the canonical tester T(S) may deadlock with I before reaching a correct final state. We put into evidence the role of the equivalence relation, conf-eq, associated naturally with conf. An important result states that if S1 conf S2, their canonical testers T1 and T2 must also satisfy T1 conf-eq T2, and reversely. Therefore, the best approach is to define the canonical tester modulo conf-eq, whereas it is currently defined modulo the testing equivalence te. Taking into account that conf-eq is weaker than te, we were able to propose a minimum canonical tester which is simpler than T(S): unlike T(S), it may have fewer traces than the specification S. The term minimum means that no trace from this tester can be deleted without losing the exhaustive test property or, stated otherwise, without taking the risk of accepting an invalid implementation (in the conf sense). [less ▲] Detailed reference viewed: 18 (1 ULg) On the role of implementation relations in the design of distributed systems using LOTOSLeduc, Guy ![]() Book (1991) Detailed reference viewed: 29 (5 ULg) Relations d'implémentation et transformations autorisées d'une spécification LOTOSLeduc, Guy ![]() in Réseaux et Informatique Répartie (1991), 1 Detailed reference viewed: 8 (1 ULg) Les applications informatiques des années 90 et leurs implications sur les réseaux de communicationLeduc, Guy ![]() in Bulletin Scientifique de l'Association des Ingénieurs Electriciens sortis de l'Institut Electrotechnique Montefiore (1991), 2 Detailed reference viewed: 12 (2 ULg) A Framework for the Taxonomy of Synthesis and Analysis Activities in Distributed System Design; ; et al in Speth, R. (Ed.) Research into Networks and Distributed Applications (1988, April) Detailed reference viewed: 15 (0 ULg) Architectural and Behavioural Modelling in Computer Communication; ; et al in Barton, M.; Dagless, E.; Reijns, G. (Eds.) Distributed Processing (1987, October) Detailed reference viewed: 16 (3 ULg) The Intertwining of Data Types and Processes in LOTOSLeduc, Guy ![]() in Rudin, Harry; West, Colin (Eds.) Protocol Specification, Testing and Verification, VII (1987, May) Detailed reference viewed: 4 (1 ULg) LOTOS, un outil utile ou un autre langage académique ?Leduc, Guy ![]() in Danthine, André (Ed.) Les réseaux de communication - nouveaux outils et tendances actuelles (1987, January) Detailed reference viewed: 27 (2 ULg) Assessing the Service Provided by a Connection-less ProtocolLeduc, Guy ![]() in Diaz, Michel (Ed.) Protocol Specification, Testing and Verification (1985, June) Detailed reference viewed: 7 (2 ULg) |
||