Programming for XML

There are many emerging applications for XML. Although there are many tools availalbe, an open question is the right programming paradigm to process XML data. Today, the most popular solutions are based on extensions to existing programming languages (e.g., Java, Python or PHP) with XML-specific libraries and APIs. Such libraries either represent the XML data as a virtual tree, or they read the XML data in a streaming (push or pull) fashion. This approach has the obvious problems that arise from the impedance mismatch between the XML type system and the type system of the host language. Moreover, the code written in such programming languages cannot be (easily) optimized using traditional techniques; good performance, scalability, and service-level guarantees is difficult to achieve for such programs on large datasets. Recently, several proposals for new programming languages have been made in both industry and the research community. One prominent example is Microsoft's XLinQ language. Another prominent example of XML processing in Web-based applications is AJAX (Asynchronous Java Programming with XML). In academia, XL, XStatic, Links, and several other languages have been proposed. All these solutions follow different philosophies and address critical design questions in different ways. This tutorial gives an overview of the current generation of programming languages for data-intensive XML applications. Furthermore, this tutorial compares the possible solutions based on a few comparative practical criteria. The tutorial shows how each solution addresses the design questions in different ways and gives the tradeoffs in terms of capabilities and optimizability of these languages are.