DEBVisDic – First Version of New Client-Server Wordnet Browsing and Editing Tool Aleˇ s Hor´ ak and Karel Pala and Adam Rambousek and Martin Povoln´ y Faculty of Informatics, Masaryk University Brno Botanick´a 68a, 60200 Brno, Czech Republic E-mail: {hales,pala,xrambous,xpovolny}@fi.muni.cz Abstract In this paper, we present the new wordnet de- velopment tool called DEBVisDic. It is built on the recently developed platform for client- server XML databases, called DEB ii. This platform is able to cover many possible appli- cations, from which we concentrate on the new, complete reimplementation of one of the most- spread wordnet editor and browser – VisDic. We argue for the benefits the new DEB ii plat- form brings to wordnet editing and to XML databases in general. In the paper, we describe the state of the implementation, the insides and interfaces of the DEBVisDic tool. We also dis- cuss its functionality and some distinctions in comparison with other dictionary writing sys- tems. 1 Introduction The original wordnet, Princeton WordNet, is one of the most popular lexical resources in the NLP field (Fellbaum, 1998). It was followed by multilingual EuroWordNet 1, 2 projects (1998- 99) (Eur, ) and Balkanet project (2001-4) (Bal, ) in which the wordnets for 13 languages have been developed (English, Dutch, Italian, Span- ish, French, German, Czech, Estonian, Bulgar- ian, Greek, Romanian, Serbian and Turkish). In the course of this work the software tools for browsing and editing wordnets have been designed and implemented, without whose the job could hardly have been performed. The first browser developed at Princeton is still used there and can be found at http://www.cogsci. princeton.edu/~wn/ or charity.princeton. edu, within the EuroWordNet project the Po- laris (and Periscope) tools have been imple- mented and used (M., 1998). For Balkanet project the browser and ed- itor VisDic (works both on Windows and Linux platform and can be downloaded from the page http://nlp.fi.muni.cz/projects/ visdic/) has been prepared at the NLP Lab- oratory at the Faculty of Informatics Masaryk University (Hor´ak and Smrˇ z, 2003) since the de- velopment of the Polaris tool has been closed by 1999. In comparison with the previous tools Vis- Dic exploits XML data format thus making the wordnet-like databases more standard and ex- changeable. Not only that, thanks to the XML data format used and to its dictionary specific configurability VisDic can serve for developing various types of dictionaries, i.e. monolingual, translational, thesauri and multilingually linked wordnet-like databases. The experience with the VisDic tool during Balkanet project has been positive (Hor´ak and Smrˇ z, 2004) and it was used as the main tool with which all Balka- net wordnets were developed. VisDic, however, has its disadvantages, par- ticularly it is not based on the client/server ar- chitecture and it does not allow to associate various attributes with literals and handle the links between them. It can work with links only between synsets which is a limiting feature for enriching wordnets with various sorts of infor- mation, e.g. in Czech with word derivation rela- tions existing within one part of speech as well as across them. The experience with VisDic has led us to more systematic research into the usage of XML data formats within the field of the computa- tional lexicography. In parallel, we also pay at- tention to the relations between wordnets and Semantic Web. This interest gives us a strong motivation for studying the properties of the XML data formats and tools for working with them. Thus we set as our task to design and imple- ment a more universal dictionary writing system that could be exploited in various lexicographic applications to build large lexical databases. The system has been called Dictionary Editor and Browser (further DEB) (Smrˇ z and Povoln´ y,