By: 	
	Kinga Laura Dobolyi
	University of Virginia, Department of Computer Science
	dobolyi@virginia.edu
	http://www.cs.virginia.edu/~kld5r/
	
	Appears in "Modeling Consumer-Perceived Web Application Fault Severities for Testing" by Kinga Dobolyi 
	and Westley Weimer, International Symposium on Software Testing and Analysis. July 2010.
	http://www.cs.virginia.edu/~kld5r/issta2010.pdf
	
	
README:
	Requirements (for running in UNIX environment):
		- paper3Web.sh (the shell script that processes all of your input HTML files)
		- AutomaticSeverity.java (this file)
		- Tidy.jar (used in this class, needs to be on your classpath)
		- parseBothPaper3.py (creates a mapping between two HTML files)
		- html5lib (python library)
		- xml (python library)
		- file.txt (lists all the input file pairs to be processed)
		- faults/... (the folder of the input file pairs)
		
	All of the necessary files and libraries come shipped with the tool. If you unzip AutomaticSeverity.tar, 
	and therefore, if you have all of these files in the same directory, you can run the tool as:
		%sh paper3Web.sh
		
	paper3Web.sh expects a file called file.txt that contains the names of all the input files
	you with to pass to this tool. The input files must be in a subfolder called "faults" and
	must end in ..._correctHTML.html and ..._incorrectHTML.html in order to be processed. 
	See the example correct and incorrect file provided (in the faults folder).
	
OUTPUT:
	This file will output results to a file called automatic.csv listing the severity rating
	under the column SCORE, followed by feature values.
	
	The rest of the output is generated by the parsing libraries and can be ignored. Note that
	the parsing libraries may take up to 30 seconds to parse files, and will time out afterwards.