Xcerpt and visXcerpt: From Pattern-Based to Visual Querying of XML and Semistructured Data

With the advent of XML as a format for data exchange and semistructured databases, query languages for XML and semistructured data have become increasingly popular. Many such query languages, like XPath and XQuery, are navigational in the sense that their variable binding paradigm requires the programmer to specify path navigations through the document (or data item). In contrast, some other languages – such as UnQL [1] and Xcerpt [2] – are pattern-based: their variable binding paradigm is that of mathematical logics, i.e. the programmer specifies patterns (or terms) including variables. Arguably, a pattern-based variable binding paradigm makes complex queries much easier to specify and to read, thus improving the programming efficiency. Sustaining this first claim with practical examples is one of the objectives of the present demonstration. Xcerpt [2] is an experimental pattern-based query and transformation language for XML and semistructured data. Xcerpt uses patterns both for binding variables in query expressions and for reassembling the variables (bound to data items in query expressions) in so-called construct terms. Arguably, a pattern-based document construction combined with a pattern-based variable binding results in a rather intuitive, user friendly, and programming efficient language. Sustaining this second claim is another objective of the present demonstration. Xcerpt is experimental in the sense that its purpose is to investigate and test another, non-navigational approach to retrieve data from the Web than that of the widespread query languages XPath and XQuery. Nonetheless, Xcerpt has been prototypically imple-