Improving rich internet application development using patterns

With changes of customer requirements, web development, especially developing Rich Internet Applications (RIA) with complex widgets and data-driven behavior can be a time-consuming task. In our previous work [3], we have presented a test-driven web development approach using ClearScript test cases as requirements to automatically generate widgets, and thus reduce the barrier of web development and testing. We extend on this work, and develop a machine learning based algorithm to identify RIA patterns [1] from requirements specified as test cases, and automatically instantiate them using simple rules. We also present performance of our algorithm and a user study which demonstrates the viability of our approach.