Development of Generic Resources for Language Technology

Moreover, it is of uttermost importance to join the interests in industry with those in the research community. New research results must be readily available to industry in order for new techniques to be useful. Often such dissemination have been made through the development of prototypes showing the possibilities and prospects of what is the current research frontier. Such prototypes illustrate the potential, but very often the code itself has been fragile and unreliable. Another possibility is to provide a repository of frameworks and tools in the form of software code ready to be used and re-used in industry projects of today and tomorrow. This puts new demands on both design and robustness of the research software. In order to be useful the design must have a generic character, i.e. it must be possible to use in new software without too much extra effort. Furthermore, the development strategy must be chosen carefully so that robustness of code is eventually achieved. Finally, the design is preferably open and the code well documented.

For the research software itself it is, similarly, important that projects are connected with each other and with users for research systems and modules to evolve beyond a certain point into useful artifacts. In particular, we have found a need for sharing programming resources between research projects. To push the level of research software forward it is, therefore, important to stimulate interactions between usage and development.

In this project we intend to meet this new role in research with its new demands on the research community by the introduction of open source techniques in connection to our existing research on language technology. We intend to develop an open source code repository for dialogue systems development and document processing. An open source community is a powerful means for dissemination of research results and if useful modules are available, i.e. code that can be used without too much effort, it can encourage development of language technology that has previously been only available as research prototypes.