Ziegenhagen, Uwe Klinke, Sigbert Härdle, Wolfgang Karl
Year of Publication:
Papers / Humboldt-Universität Berlin, Center for Applied Statistics and Economics (CASE) 2004,40
Modern statistical computing requires smooth integration of new algorithms and quantitative analysis results in all sorts of platforms such as webbrowsers, standard and proprietary application software. Common statistical software packages can often not be adapted to integrate into new environments or simply lack the demands users and especially beginners have. With Yxilon we propose a vertically integrable, modular statistical computing environment, providing the user a rich set of methods and a diversity of different interfaces, including command-line interface, web clients and interactive examples in electronic books. This architecture allows the users to rely upon only one environment in order to organize data from a variety of sources, analyse them and visualize or export the results to other software programs. The design of Yxilon is inspired by XploRe, a statistical environment developed by MD*Tech and Humboldt-Universität zu Berlin. Yxilon incorporates several ideas from recent developments and design principles in software engineering: modular plug-in architecture, platform independence, and separation of user interfaces and computing engine.