Predicting Source Code Irregularities in Automated Code Conversion Systems Using SRASG
With newer technologies emerging, conversion of legacy code to accommodate newer generation hardware and operating systems is essential. In legacy systems the dependency on the platform for which it is designed is higher. Major tasks of a software engineer are to continually alter code to prevent it becoming obsolete. Work has been extensively done to automate code generation from one language to another. However it is found not all parts of the code is converted successfully with larger lines of code more error prone compared to smaller chunks of code. In this paper it is proposed to implement a novel slicing method to break down the given legacy module to smaller module and propose a data mining method to identify source codes which may be erroneously converted.