Don't regenerate random sample of web page urls and full web page url listing files if they already exist, as we don't want to accidentally lose any manual work done to the random sample file when rerunning the code. 2. Fixed up duplicate logging of command run by SafeProcess by recompiling. 2. System output commands replaced by logger calls where appropriate, since logger is also outputting to console. 4. Clarifying some log4j.props comments