Mireille Clerbout

Synchronization languages

Synchronization expressions introduced in [7] within the framework of the Pm-C project are a high-level construct which allow a programmer to express minimal synchronization constraints of a program in a distributed context. The study and the implementation of these expressions are based on their associated synchronization languages. Synchronization languages have been introduced in [9, lo]. They give a way to implement synchronization expressions and specify their semantic. These languages represent distributed systems whose behaviour respects synchronization constraints expressed by the programmer with synchronization expressions. So these languages describe all the correct executions of a program. In [9, lo] Guo, Salomaa and Yu propose a characterization of synchronization languages based on a rewriting system named R which generalizes partial commutations. This system gives a way to rewrite a word representing a parallel execution into a word with a lower or equal degree of parallelism. Guo et al. show that every synchronization language is closed under the system R and they conjecture that it is sufficient for a regular language to be closed under R to be a synchronization language. We show that the conjecture is true in the particular case of languages expressing the synchronization between two distinct actions. We also show that the conjecture is false in the general case. @ 1999-Elsevier Science B.V. All rights reserved

About Synchronization Languages

Synchronization languages are a model used to describe the behaviors of distributed applications whose synchronization constraints are expressed by synchronization expressions. Synchronization languages were conjectured by Guo, Salomaa and Yu to be characterized by a rewriting system. We have shown that this conjecture is not true. This negative result has led us to extend the rewriting system and Salomaa and Yu to extend the definition of synchronization languages. The aim of this paper is to establish the link between these two extensions, we show that the behaviors expressed by the two families of synchronization languages are only separated by morphisms.

Partial Characterization of Synchronization Languages

Synchronization languages are associated with synchronization expressions, a high-level construct which allows a programmer to express synchronization constraints in a distributed context. We give a negative answer to a conjecture enunciated by L. Guo, K. Salomaa and S. Yu which aims at characterizing synchronization languages in terms of regular languages closed under a rewriting system. Then we propose an extension of the system which gives a positive answer for a class of regular languages.

Decomposition of semi commutations

We define atomic semi commutations as being associated to independance relations the form of which is A × B, in which A and B are two disjoint subsets of the alphabet. We prove that semi commutations can be decomposed in weaker semi commutations if and only if they are not atomic. We then deduce that every semi commutation can be obtained by a composition of atomic semi commutations and we suggest a decomposition algorithm.

Synchronization Languages and Rewriting Systems

We extend the rewriting system defined by Guo et al. in order to characterize closure properties of synchronization languages. The extension is shown to have well-known properties of commutation systems and to be the rewriting system which suits the synchronization languages best. We give a characterization of finite synchronization languages in terms of languages closed under the extension.

Semi-commutations and Partial Commutations

The aim of this paper is to show that a semi-commutation function can be expressed as the compound of a sequential transformation, a partial commutation function, and the reverse transformation. Moreover, we give a necessary and sufficient condition for the image of a regular language to be computed by the compound of two sequential functions and a partial commutation function.

Generalized Synchronization Languages

Generalized synchronization languages are a model used to describe the behaviors of distributed applications whose synchronization constraints are expressed by generalized synchronization expressions -- an extension of synchronization expressions. Generalized synchronization languages were conjectured by Salomaa and Yu to be characterized by a semi-commutation. We show that this semi-commutation characterizes the images of generalized synchronization languages by a morphism-like class of rational functions.