Manual testing is a tiring job and when it comes to large software products, it better be not talked of. According to the recent studies, software developers have been found investing a lot of time into the bug discovery and resolution, making it an expensive task. However, engineers have seen the rise of automated testing, the cheaper counterpart, but where do the required test cases come from? This is exactly the point raked up by Andreas Zeller, professor of Software Engineering at #-Link-Snipped-#. But Zeller, with other two computer scientists Nikolas Havrikov and Matthias Hoschele, seems to have developed a solution for generating the test cases and processing them as well.


The software named XMLMATE works by automatically generating the test cases and using them for testing the given piece of code. As the input given is used to generate initial set of cases, hence it should be structured in a specific way. The XML inputs are parsed into a central structure called tree. Collections of these trees, termed as chromosomes are provided to the genetic algorithm. This algorithm is responsible for the generation of different test cases. The testing of input is modeled on biological evolution, hence, the input that is processed by the maximum units of code survives. Havrikov who implemented XMLMATE admits that finding the errors is a tough task and the real challenge lies in maximum code coverage. according to the developers of the system, the generator mutates, recombines, generates and evolves the input to produce cases.

XMLMATE is developed in java programming language. The inputs are defined in XML and presence of XML schema could be an advantage. However, if the programming input is not XML supported, it can be configured accordingly by using various tools. The software is already on the open source and is being used by many developers. The creators believe that with their framework they can test various spheres ranging from computer networks to car sensors.

Source:#-Link-Snipped-#