An Ontological Approach to the Development of Analytical Platform Language Toolkits

The development of complex software systems is impossible without using modern modeling tools. At the design process, various models are developed: when solving each task, the attributes of processes and systems, which are significant for solving this task, are to be presented in the model. Developing analytical systems for data-intensive areas has specifics, which determine new requirements for the system functionality and architecture (interoperability, adaptability, etc.). These requirements can be implemented in the system based on a combination of two approaches: knowledge-driven development and model-driven development. This article presents an approach to creating a knowledge-driven analytical platforms, which integrate language toolkits allowing to create "on the fly" new domain-specific languages (DSLs). DSLs provide "user interfaces" which are customizable to the specifics of the tasks, solved by users with modeling tools, and to the corresponding users' domains. The architecture of the analytical platform, the graph model, and the metalanguage which are the basis of the language toolkits implementation are described in the paper. The multifaceted ontology, which is the core of the analytical platform, is presented too.