A multi-software integration platform and support for multimedia transcripts of language

Using and sharing multimedia corpora is a vital feature for research about language, but the number of different and often not easily compatible tools available makes this difficult to do. As the aims of the COLAJE project are to use multimodal linguistic data about language development in oral and sign languages, it was necessary to create a system (VICLO) that allowed sharing and using data coming from at least three different sources Clan (CHILDES), Elan (MPI) and Praat (U. of Amsterdam). For this reason, a multipurpose storage format based on the TEI was created, which allowed us to store information coming from all (these) origins, and include every type of specific information. When part of the information is processed by a specific software, the changes are integrated later in the system without loosing information specific to other software. Thus it is possible to store information shared and not shared between the different corpus editing tools. This common base allowed us to implement complementary features such as fine-grained participant and metadata information, common visualisation and data-retrieval tools. VICLO is based on XML technology and all data can be displayed using all purpose web browsers.