After describing the SDL current position within the modelling languages domain, this document describes the route plan to 2010 version of the language.

"A real time / embedded software is basically a software nobody sees but it is the one that has most constraints in the software industry such as high-level of reliability, small memory footprint, and high performance. In the same time embedded software requirements are getting more and more complex because all devices should now be able to communicate with each other and also because hardware components prices are dropping allowing the usage of powerful processors for small systems. What used to be a few hundred bytes of code in assembly language is now hundreds of thousands of lines in C code. Therefore it is a priority for all real time software developers to make sure the software architecture is good, the interfaces with other modules is well defined, the code is legible and re-usable. A reduced time to market also brings the need to develop software before the hardware is done: simulating and prototyping the software is now a must."

"A new approach for Earthquake Early Warning Systems (EEWS) is presented that uses wireless, self-organising mesh sensor networks. To develop the prototype of such IT-infrastructures, we follow a model-driven system development paradigm. For modelling a technology mix of SDL/ASN.1/UML/C++ is used to generate the code for different kind of simulators, and for the target platform (several node types). This approach is used for realizing a prototype-EEWS developed within the EU project SAFER (Seismic eArly warning For EuRope) in cooperation with the GeoForschungszentrum Potsdam."

"This work presents some new optimization approaches to implementation of Medium Access Control (MAC) layer of IEEE 802.11 wireless networking protocol using general purpose DSP and gate array systems. Optimization starts at design level. The hardware/software partitioning of the MAC’s architecture is optimized in the sense of minimal implementation burden, while maintaining the system’s functionalities and performance. The proposed partitioning and implementation technique obviates the use of any Real Time Operating System (RTOS), which leads to a simple, high speed, and low memory structure of the MAC’s software. Also, solutions such as using hash tables and pipeline processing are given and employed to gain a higher speed. The software section is implemented on the popular, low price DSP of TI’s C54 processor, whereas the hardware implementation is realized using Virtex2vp30 from Xilinx."

"Implementation of TTCN-3 in RTDS tool raised some interrogations. The concept of snapshot in the TTCN standard is not present in traditional RTOS. This paper will first make a short presentation of TTCN-3. It will then explain the alternative rationale and the concept of snapshot. A reminder of event-driven concepts in major RTOS will be done, to finish with a presentation of several solutions to implement the snapshot."

"Abstract: The TASTE tool-set results from spin-off studies of the ASSERT project, which started in 2004 with the objective to propose innovative and pragmatic solutions to develop real-time software. One of the primary targets was satellite flight software, but it appeared quickly that their characteristics were shared among various embedded systems. The solutions that we developed now comprise a process and several tools ; the development process is based on the idea that real-time, embedded systems are heterogeneous by nature and that a unique UML-like language was not helping neither their construction, nor their validation. Rather than inventing yet another “ultimate” language, TASTE makes the link between existing and mature technologies such as Simulink, SDL, ASN.1, C, Ada, and generates complete, homogeneous software-based systems that one can straightforwardly download and execute on a physical target. Our current prototype is moving toward a marketed product, and sequel studies are already in place to support, among others, FPGA systems."

"La production de systèmes critiques (contrôle de véhicule, détection de fautes, drones, etc.) requiert le respect de nombreuses exigences: ils opèrent dans des environnements contraints (domaine avionique, spatial, militaire) et s'exécutent sur des plateformes embarquées ayant des ressources limitées (capacité de calcul, taille mémoire). De plus, une erreur dans leur implémentation peut avoir de lourdes conséquences (abandon d'une mission, perte de vie) si bien que leur code doit être exempt de bogue. La conception de tels systèmes demande donc un processus de développement rigoureux, s'appuyant sur des technologies détectant tout potentiel vecteur d'erreur. Cet article présente une chaîne d'outils implémentant de tels systèmes au travers d'un cas pratique l'implémentation d'un drone d'exploration avec Linux."

"Nous allons passer en revue les concepts du langage SDL, son historique, et évaluer les forces et faiblesses respectives de SDL et UML afin de projeter ce qui, à notre avis, serait une solution optimale pour l’utilisation de l’un ou l’autre de ces langages."