Web application internationalization and localization in action
Terence Parr · 2006
A template engine that strictly enforces model-view separation has been shown to be at least as expressive as a context free grammar allowing the engine to, for example, easily generate any file describable by an XML DTD. When faced with supporting internationalized web applications, however, template engine designers have backed off from enforcing strict separation, allowing unrestricted embedded code segments because it was unclear how localization could otherwise occur. The consequence, unfortunately, is that each reference to a localized data value, such as a date or monetary value, replicates essentially the same snippet of code thousands of times across hundreds of templates for a large site. The potential for cut-and-paste induced bugs and the duplication of code proves a maintenance nightmare. Moreover, page designers are ill-equipped to deal with code fragments. But the difficult question remains: How can localization be done without allowing unrestricted embedded code segments that open the door to model-view entanglement? The answer is simply to automate the localization of data values, thus, avoiding code duplication, making it easier on the developer and designer, and reducing opportunities for the introduction of bugs--all-the-while maintaining the sanctity of strict model-view separation. This paper describes how the StringTemplate template engine strictly enforces model-view separation while handily supporting internationalized web application architectures. Demonstrations of page text localization, locale-specific site designs, and automatic data localization are provided.