Collaborative editing of Multimodal Annotation Data

The annotation of multimodal speech corpora is a particularly tedious task, since annotatable events can be composed of smaller events that span across several modalities (e.g. speech and gesture), which imposes the need to operate on the same data, using a wide range of different tools in order to cover all the different modalities and layers of abstraction within multimodal data. MonadicDom4J has been developed as a highly generic general purpose java-based Rich Client framework that opens the possibility to simultaneously operate on any kind of XML data through several different views and from several remote locations. It allows for the dynamic allocation of plugins, needed to render a given type of XML markup, and takes care of the concurrency between different sites viewing the same data, by means of differential synchronization. The demonstration will involve several different applications ranging from general textual hyperdocument editing to multimodal annotation tools, whose contents can be freely intermixed, interlinked and transcluded into different contexts, using drag and drop interaction. The audience will have the opportunity to try collaborative editing on the presented examples from their own devices.