High Energy Physics Libraries Webzine Issue 8 / October 2003 Experience in Computer-Assisted XML-Based Modelling in the Context of Libraries Marko Niinimäki, Vesa Sivunen (*) Abstract: In this paper, we introduce a software called Meta Data Visualisation (MDV) that (i) assists the user with a graphical user interface in the creation of his specific document types, (ii) creates a database according to these document types, (iii) allows the user to browse the database, and (iv) uses native XML presentation of the data in order to allow queries or data to be exported to other XML-based systems. We illustrate the use of MDV and XML modelling using library-related examples to build a bibliographic database. In our opinion, creating document type descriptions corresponds to conceptual and logical database design in a database design process. We consider that this design can be supported with a suitable set of tools that help the designer concentrate on conceptual issues instead of implementation issues. Our hypothesis is that using the methodology presented in this paper we can create XML databases that are useful and relevant, and with which MDV works as a user interface. 1 Introduction The aims of this paper are (i) to survey modelling from the point of view of computer science and information studies, (ii) to study how XML can be used in this context and what its benefits are, and (iii) to outline an implementation of an XML-based modelling tool. The details of the implementation will be explained in a forthcoming paper. According to Reingruber and Gregory [1 ], a model in general is 'a hypothetical or stylized representation' it attempts to capture in meaningful form some larger, or in some cases smaller, object that exists or will exist. Modelling, naturally, is the process of forming these models. We call the object of modelling a domain of interest or, briefly, domain. There are possibly dozens (if not hundreds) of different accounts of modelling in different domains - engineering, economics, medicine, and social sciences, to name just a few. In the scope of this paper we limit our interest to modelling in computer science and information studies, and in order to do so, we make some simplified assumptions: • Assumption 1: models and modelling are important tools for capturing information about the domain. • Assumption 2: this information can be expressed in textual form, that we call documents. • Assumption 3: the structure of the documents reflects the features of the domain.