Automatic Detection of Failures and Anomalies Using Metamorphic Relations for Search Engines
Author(s):
K.Sri Sandhya, S. Vasundra
Keywords:
Software quality, verification, validation, quality assessment, oracle problem, lack of system specification, metamorphic testing, user-oriented testing, search engine
Abstract
Software Testing is an essential and costly activity in software development to detect faults in programs. An oracle is a mechanism by which testers can decide whether output of program under testing is correct or not. In some cases, it is too difficult to find oracle which in turn causes oracle problem. Metamorphic Testing (MT) method is used in such cases to test programs without involvement of an oracle. In the existing system metamorphic testing has been extended into a quantifiable approach for software quality assessment which includes verification and validation of software correctness for 4 different web search engines: Google, Bing, Chinese Bing and Baidu. The system detects various kinds of failures and also operational profiles having a significant impact on quality of search. In the present system Metamorphic Testing is further extended on the search engines by implementing Metamorphic Relations which automatically detects the failures and anomalies. It also provides various hints for the construction of run time self-correction mechanism which will be very helpful for the user to get exact information from the search engine.
