International audienceWhen a project is realized in a globalized environment, multiple stakeholders from different organizations work on the same system. Depending on the stakeholders and their organizations, various (possibly overlapping) concerns are raised in the development of the system. In this context a Domain Specific Language (DSL) supports the work of a group of stakeholders who are responsible for addressing a specific set of concerns. This chapter identifies the open challenges arising from the coordination of globalized domain-specific languages. We identify two types of coordination: technical coordination and social coordination. After presenting an overview of the current state of the art, we discuss first the open challenges arising from the composition of multiple DSLs, and then the open challenges associated to the collaboration in a globalized environment

Invited talk abstractThese last five years have seen the generalization of model usage, as their adequation to the design of complex systems began to be more and more understood, and researched. Modeling using specifically tailored models is now common practice. Modeling tools, especially in Eclipse, matured a lot and they are fit for a lot of different contexts and requirements; from the French unemployment office to the NASA. This rising success is not complete yet though, and collaborative modeling is barely taking off, holding back wider adoption.Let's go back all the way to 2006. At that time, during the Eclipse Summit Europe conference, the community at large voices its concerns : there is no available open-source component aiming at comparing and merging models; while such a tool is critical to the success of all modeling-related technologies. Models were becoming more and more central in the development process, their importance rising to the level of the source code's... and yet there was no tool designed to allow teams to work on models collaboratively. It is in light of this observation that we decided to start the EMF Compare project.The challenges were then to be able to compare any kind of model, be it specific or based on a standard, with a good scalability and accuracy of the matching algorithms; and have this comparison integrated with the many different version control management systems available. At that time, there was but a few commercial tools barely starting to provide basic support for collaborative work on models; and the community recognized our commitment in providing a reliable open-source stack.Five years later, both technologies and practices have dramatically evolved. What seemed obvious in 2006 turned out to be but one option among many others, and as time went by we have seen that the original challenges only were a subset of what we were really going to face. This talk will start with the story, from the perspective of an Eclipse commiter and Obeo insider, of what happened during those five past years : how practices and requirements evolved, how our component adapted to these changes and has been used in practice, how we tackled collaborative modeling needs together with our adopters... I will then move on to what is going on right now from the Eclipse side; focusing on how EMF Compare, along with -among others-Mylyn and CDO, are playing a major role in settling the basis of a sound modeling platform supporting collaboration.