Edite - A Natural Language Interface to Databases A new dimension for an old approach Paulo Reis João Matias Nuno Mamede INESC Av. Duque d’Ávila 23, 1000 Lisboa, Portugal {Paulo.Reis, Joao.Matias, Nuno.Mamede @inesc.pt} Abstract This article presents the Edite system, a Natural Language Interface for Databases (NLIDB), that tries to explore the advantages of joining natural language processing with the expressiveness of graphical interfaces. In order to guarantee a permanent adaptation of this type of solution to a dynamic domain one should consider two critical fundamental factors: extensibility and portability. An overview of the system architecture is presented, emphasising those choices that were imposed by the demands of portability and extensibility. Several general problems of natural language processing that were faced in constructing the system are discussed. Future work is highlighted. Keywords: tourism, natural language processing, intermediate representation languages, database querying. 1 . Introduction The importance of the tourist industry for the Portuguese economy has significantly increased in the last 10 years. It accounts nowadays for approximately 8% of the GDP and equals that of the financial sector. The main reasons for the success of tourism are: the climate, the historical and cultural heritage, the tourist infrastructure, the hospitality of the Portuguese people and the relative close emissary markets. It is important to create conditions to consolidate the registered growth, which can be achieved through a firm development strategy focusing on the quality of supply and human resources and on diversification of markets and products. National tourism has adopted new instruments in order to reinforce their performance focusing on the new information and communication technologies. The Inventory of Tourist Resources (IRT) emerges as the largest R&D Portuguese project in this area, actually exercising fundamental support on tourist ordering and planning and on forming a global reference point our tourism. The IRT initially emerged to eliminate the shortage of institutional information, positioning itself as the largest data repository of tourist resources on Portuguese tourism. IRT’s vast variety of purposes determined the adoption of an architecture able to support the integration of a large informational universe on a stage of multiple user segments and on the various applicational objectives (e.g., planning, promotion, distribution, management …). Nowadays, IRT’s informational domain uses Internet, multimedia “kiosks”, GIS, graphic user interfaces and “natural language interfaces”. The supporting technologies on the diverse functions of IRT are in well marked, distinct stages of development, from a scientific and technological point of view as well as from a commercial point of view.