UNIVERSITY OF HERTFORDSHIRE COMPUTER SCIENCE RESEARCH COLLOQUIUM presents "Exploiting Abstract Syntax Trees to Extract Code Constructs Associated with Software Faults" Thomas Shippey (School of Computer Science, University of Hertfordshire) 17 December 2014 (Wednesday) 1 pm - 2 pm Hatfield, College Lane Campus Seminar Room D102 Everyone is Welcome to Attend Refreshments will be available Abstract: Identifying defects in code early is important. A wide range of static code metrics have been evaluated as potential defect indicators. Most of these metrics offer only high level insights into the code with no metric clearly performing best in defect prediction. Effectively identifying defect-prone code remains a challenge. In this talk I will present code snippets as fine-grained low level features of Java code and identify what affect specific code snippets have on faults. I extracted code snippets from sequences of Java 'kinds' extracted from the Java Abstract Syntax Tree. I used Binomial statistical tests to determine relationships between code snippets and faults in three open source systems. The results show that some code snippets are very significantly related to faults in some systems and the effect on faults of some snippets is large. The occurrence of some frequently occurring code snippets in a method can make that method four times as likely to contain a fault. Frequently occurring fault-prone snippets include those associated with indentifiers, method calls and variables. The talk will establish that code snippet analysis offers developers a promising approach to identifying potentially defective code. The technique also has potential uses in other areas, such as, automatic patching, code cloning and defect prevention. ------------- Hertfordshire Computer Science Research Colloquium http://cs-colloq.stca.herts.ac.uk