The BLARK concept and BLARK for Arabic

The EU project NEMLAR (Network for Euro-Mediterranean LAnguage Resources) on Arabic language resources carried out two surveys on the availability of Arabic LRs in the region, and on industrial requirements. The project also worked out a BLARK (Basic Language Resource Kit) for Arabic. In this paper we describe the further development of the BLARK concept made during the work on a BLARK for Arabic, as well as the results for Arabic. 1. The BLARK concept The BLARK defines, ideally in a language independent way, the minimal set of language resources to do any precompetitive language and speech technology research at all for a language. After the first BLARK article (Krauwer, 1998) in the ELRA Newsletter, the idea was taken up by the Dutch Language Union (DLU). Daelemans & Strik (2002) give an overview of the steps taken by DLU to define the contents of the BLARK for Dutch and to assign priorities. Unfortunately this document is only available in Dutch. A summary of the work of the DLU and the results of the Dutch BLARK exercise can be found in Binnenpoorte et al (2002). Later the ENABLER project also contributed to the definition of the BLARK concept. The starting point of the definition process in Binnenpoorte et al were 8 classes of applications, seen as being the most relevant application categories at that moment: computer assisted language learning, access control, speech input, speech output, dialogue systems, document production, information access and translation. For each of them it was established which modules would be needed to make them (e.g. morphological analysis, text to phoneme converter), and for each of these modules it was analyzed which language data (e.g. data sets, descriptions) they would require, as well as their relative importance. The results were put together in a large matrix, on the basis of which one can determine which components serve most applications, and which data are most needed for most applications, i.e. which elements should be part of the BLARK. NEMLAR took this as the point of departure. We distinguish the BLARK definition which is the general concepts governing the BLARK, and the BLARK specification which is its instantiation for a given language, here Arabic. For the general discussion of the BLARK concept we first discuss a few important issues: availability, quality, quantity and standards.